An Exploration of Beliefs about Gender Differences in Language Use

It is the natural order of things for humans to acquire beliefs and conform to stereotypes in an attempt to explain phenomena surrounding them. These mental constructs are known to have a pervasive influence on the way people think and act, and therefore are partly responsible for shaping our social reality. Thus, due to their impact, scientific exploration is needed to illuminate their nature and so enable humans to act upon these findings. Beliefs or stereotypes that are being studied in this particular research are those held about the differences in language use by men and women. Acknowledging that people in Bosnia and Herzegovina largely comply to traditional, patriarchal social norms, this study aims to elucidate the matter by investigating whether students of a private university situated in Sarajevo, Bosnia and Herzegovina, conform to widespread stereotypes about language and gender, women’s speech and men’s speech in particular, and whether males and females differ in conformity to the stereotypes.

With the emergence of gender studies in the 1970s, the tendency to explore and understand the differences between genders found its place within various scientific disciplines. The study of language and gender is one such interdisciplinary field, which aims to provide answers to the question whether men and women talk differently. Asking such a question is considered to be of paramount importance for understanding language as a social practice that influences the way humans perceive and position themselves and others in social schemes. The task of such scholarship entails "the broader task of interrogating and transforming existing conceptual schemes" (McElhinny, 1998, p. 21), challenging the assumptions commonly underlying gender-based research. These assumptions, largely conceived from stereotypes rather than on solid scientific evidence and firmly grounded on entrenched beliefs about the manner in which women and men use language adhered to since childhood (Coates, 2013), however, constitute a principal part of so-called 'folk knowledge'.
As much as it is important to study gender differences in linguistic usage through the analysis of actual language data, a more detailed insight into speakers' attitudes towards these differences might be of key importance for further understanding of human perception of social reality through language, and thereby for the development of the field. The present study thus attempts to shed some more light on this issue by exploring beliefs International Burch Esma Latić, E., & Čeljo, A.B. (2018). An Exploration of Beliefs about Gender Differences in Language Use. Journal of Language and Education,4(3), 48-57. doi:10.17323/2411-7390-2018-4-3-48-57 University students have about gender differences in linguistic usage. More specifically, it aims to discover, firstly, if the stereotypes are present among students and, secondly, whether male and female students significantly differ in terms of their conformation to common stereotypes about the way women and men use language.

Literature Review
One of the first major scholars to propose a theoretical construct of the way women speak, differing from the way men do, was Robin Lakoff, who sparked and popularized the discussion with her book, Language and Women 's Place (1975). Namely, Lakoff (1975) claims that "women's language" is a form of language somehow deficient compared to the "neutral" form of language -the one the author uses to refer to the form of language men use. As stated in Lakoff (1975), "women's language" is taught to women in their childhood and expected to be used by them in adulthood at the expense of their own expressiveness. Thus, women are incapable of expressing themselves fully and genuinely because of the limitations of the language that have been imposed upon them by society, which leads to perpetual social gender inequality (p. 40). "Women's language", according to Lakoff (1975), is characterised by the following features: (1) vocabulary related to their interests (e.g. magenta, shirr), (2) "empty" adjectives (e.g. adorable, charming), (3) interrogative intonation where declaratives are expected (e.g. "What's your name, dear? Mary Smith?"), (4) hedges (e.g., I think, sort of, I wonder), (5) intensifiers (e.g., so, very), (6) hypercorrect grammar (e.g. ending 'g' is almost never dropped out in pronunciation of words, such as "singing" becoming '"singin"), (7) excessively polite forms (e.g., using "please" and "thank you" often), (8) lack of sense of humour, and (9) speech in italics (using tone to emphasize certain words while speaking) (p. 78-81). Lakoff's work sparked a widespread interest in the unexplored area examining whether men and women do speak differently (e.g. Amir et al., 2012;Brown, 1998;Bucholtz, 2004;Cameron, McAlinden, & O'Leary, 1989;Coates, 2003;De Klerk, 1992;Dubois and Crouch, 1975;Fishman, 1980a;Gibson, 2009;Gomm, 1981;Holmes, 1984Holmes, , 1987Irwin, 2002;Kuha, 2005;Livia, 2004;McConnell, 2004;McElhinny, 2004;Mei, 2006;Nemati and Bayer, 2007;Preisler, 1986;Schleef, 2009;Uhlman, 2015) and it encompassed the core of the stereotypes of women's speech that still seem to exist in different quantities and qualities across the world's societies. However, it also faced intense criticism for its lack of solid scientific evidence, as well as the reference to "men's language" as the norm (e.g. Coates, 2013;Crawford, 1995;Holmes, 1987;Leaper & Robnett, 2011;O'Barr & Atkins, 1980;Spender, 1980). Coates (2013) also maintains that, apart from an academic audience, the general public still seem to adhere to "a world view where women should learn to be more like men" (p. 7) and "grow up to believe that women talk more than men, that women 'gossip', that men swear more than women, that women are more polite, and so on" (p. 86). The author further argues that beliefs that women's vocabulary is filled with words of less significance for the language (p. 10), that "women's language" is much more polite and sophisticated, or simply more "ladylike" (p. 15) and pervaded by verbosity have persisted throughout centuries (Coates, 2013). Despite the fact that there is no clear evidence for such claims, still, as maintained in Coates (2013), "Western European culture is imbued with the belief that women do talk a lot, and there is evidence that silence is an ideal that has been held up to (and imposed on) women for many centuries" (p. 26).
Numerous studies that tested Lakoff's claims refuted the idea that there is a uniform difference between males and females in their linguistic practice which reflects their social inequality and asserted that the differences that possibly exist do not necessarily reflect the powerlessness Lakoff talks about when discussing "women's language" (e.g. Brown, 1998;Cameron et al., 1989, De Klerk, 1992Dubious & Crouch, 1975;Eckert & McConnell-Ginet, 2003;Holmes, 1995;Nemati & Bayer, 2007;O'Barr & Atkins, 1980). However, various other studies seem to confirm the existence of the aforementioned entrenched beliefs and stereotypes about women's speech. For instance, in 1967, when Lakoff's books still had not been published, in a study of blue-collar families and their gender stereotypes, Komarovsky found that both genders were aware of the widespread perception of "women's language". One of the female participants said: "[Men] think we [women] are silly and talk too much. They think that women gossip a lot and they are against it", while a male participant commented that women "talk about kidstuff and trivia like Mrs. X had her tooth pulled out" (as quoted in Haas, 1979, p. 619). They also agreed that the topics usually discussed by women are family and social problems, while men like to talk about politics, cars, sports, work, etc. Klein (1971) inferred that both men and women talk primarily about their work and secondly about the things that are of interest to them (p. 73). Moreover, Garcia-Zamor (1973) investigated the stereotypes in pre-school children by asking girls and boys to judge whether some utterances were produced by a female or a male doll and found that stronger expletives as well as competitive and agressive language were labelled as male, and the more euphemised utterances as female (as quoted in Haas, 1979, pp. 617-622). Kramer (1974) used captions from the New Yorker's cartoons and asked college students to judge them as male or female and asked for an explanation of their choices. For most of the captions (79%), male and female students made the same choices when guessing gender of the speaker, conforming to similar linguistic stereotypes, such as that women speak more and men are more authoritative, leading them to their conclusions. Moreover, Kramer (1977) investigated the stereotypes about the speech of females and males and found that male speech is commonly depicted as boastful, militant, authoritarian, forceful, aggressive, blunt, dominating and loud, using swear words, showing anger rather than concealing it, straight to the point, having sense of humor (pp. 156-157), whereas the typical female speech was characterized as gentle, fast, friendly, emotional, polite, smooth, open and self-revealing and enthusiastic (pp. 156-157), women are claimed to smile a lot when talking, to speak in good grammar, discuss trivial topics, enunciate clearly, use hands and face to express ideas, gossip, express concern for the listener, use many details, smile a lot when talking and use good grammar (p. 159). Siegler and Siegler (1976) tested whether college students would attribute strongly assertive forms more often to males and less assertive ones to females, as well as whether the forms assigned to men would be judged more intelligent than those associated with females. The results confirmed that strong assertions were most often ascribed to males and they were also rated more intelligent (p. 169).
Thus, it seems that a significant amount of research has been carried out to uncover the differences in male and female language use and such studies are certainly necessary in order to eliminate the abundance of empirically unproven opinions, particularly noticeable among general public. There appears to be, however, a scarcity of studies that focused on the stereotypes of female and male speech which seem to persist in spite of evident development of humankind in terms of gender equality and change in perception of gender roles and traits. The need to investigate these particular stereotypes is summarized by Kramer (1975) who asserts that the stereotypes of the way females and males speak reflect the stereotypes of general gender roles in society and result in people trying to act and interact according to the gender roles and norms prescribed by these stereotypes (p. 152).

The Present Study
In Bosnia and Herzegovina, patriarchal sentiment is still quite strong and beliefs about gender roles are still largely based on traditional, conservative notions of how women and men should behave, thus encompassing beliefs and prescriptions about the way they speak as well. Even though males and females are believed to be afforded equal opportunities to communicate (Bećirović, 2012;Rizvić & Bećirović, 2017), young girls are often expected to avoid speaking too loudly or too assertively, particularly in mixed-gender conversations, using "inappropriate" language, even in informal situations, discussing particular "male" topics, etc. in order not to be perceived as ill-mannered or "tomboy-like", while the same expectations are rarely held for young boys. The situation, naturally, varies across different social classes and urban/rural environments and is highly dependent on individuals' levels of education and self-consciousness. However, this entire spectrum of socially desirable roles perpetuates the state of gender inequality generally evident in various spheres of life in Bosnia and Herzegovina.
This particular article explores one part of society and its members might not be fully representative of the entire population. However, due to the paucity of evidence in the literature, especially in the Bosnian context, this study is an attempt to initiate the exploration of how deeply-rooted stereotypes are and how they might account for differences existing in other aspects of life. More specifically, its aim is to explore whether younger people, namely universitylevel students, conform to the common stereotypes about gender differences in language use and whether male and female participants significantly differ in their conformation to the stereotypes about women's speech, men's speech and the relationship between language and gender in general. Thus, the following hypotheses will be tested in the process:

H1:
Due to their age and the specificity of their educational milieu, we hypothesize that the participants do not fully conform to strong stereotypes about the way women and men speak and about the relationship between language and gender. H02: There is no statistically significant difference in the female and male students' perceptions of the way women and men speak. H03: There is no statistically significant difference between female and male students' perceptions of the relationship between language and gender.

Participants
Sixty-nine undergraduate level students studying in different study years at the Department of English Language and Literature at a private university in Bosnia and Herzegovina participated in the study. The language of instruction at this institution is English. All the participants were non-native speakers of the English language coming from different cultural backgrounds. A more detailed description of the participants is presented in Table 1.

Measures and procedures
The study's questionnaire was constructed based on the list of general assumptions about the way women and men speak suggested in Keith and Shuttleworth (2008). The assumptions are somewhat similar to those proposed by Lakoff (1975), which are perhaps more general in their nature and were therefore considered appropriate for this particular study. The questionnaire consists of 22 items which can be further divided into three sections: (1) stereotypes about female speech (8 items), (2) stereotypes about male speech (9 items), and (3) general beliefs about language and gender (5 items). All items were rated on a 3-point scale, ranging from disagree (1) to agree (3). In addition to this questionnaire, a general demographic survey was distributed among participants. The participants were students of the Department of English Language and Literature at International Burch University, a private university situated in Sarajevo, Bosnia and Herzegovina. The questionnaire was distributed in a written form and completed anonymously, with the participants being informed about the aim and potential benefits of the study. The average time of completion was around 20 minutes. The overall data is internally consistent as the Cronbach Alpha reliability coefficient for 22 items is .828. The Cronbach Alpha coefficients both for the subscale perceptions of women's speech (8 items) and for the subscale perceptions of men's speech is .724, and for the subscale the relationship between language and gender is .600.

Data analysis
All data were analysed using the Statistical Package for Social Sciences (SPSS version 23). In order to analyse individual items of the instrument, a descriptive analysis was run. The independentsamples t-tests were used with the aim of investigating whether there exist statistically significant differences between male and female participants in terms of the stereotypes of the way men and women speak, general beliefs about language and gender, and likelihood of "feminine" language usage. All tests of significance in this study were set at the .05 level.

Preliminary analyses
The descriptive results of the participants' perceptions of male and female speech and the relationship between language and gender are displayed in Table 2 encompassing means, standard deviations and reliability. With the aim of investigating whether there was consistency in the participants' perceptions of female speech, the data were statistically analysed. In the output presented in Table 3, the data requested for each of the items are summarized. The highest mean (M = 2.46, SD = .719) was calculated for the assertion that women ask more questions, where a large number of participants (n = 41, 59.4%) agreed with the statement. At the other end, the lowest mean (M = 2.05, SD = .855) was found in the statement that women talk too much, where 23 participants (33.3%) disagreed and 27 of them (39.1%) agreed with the idea.
The results in this section show that the participants did not categorically agree or disagree with most of the statements, but rather demonstrated high variance in responses -leading us to the conclusion that the participants generally did not strongly conform to the existing traditional stereotypes as might be expected. Major consistencies, where more than 50% of the participants either agreed or disagreed with the statements, were found in only two items; namely, 39 participants (56.5%) claimed that women are more polite and 41 participants (59.4%) said that women ask more questions. The first one is completely understandable if we take into consideration that in Bosnia and Herzegovina the social norm is that women should be more polite than men, both in their behaviour and language use. A particularly interesting finding was that a number of participants (n = 23, 33.3%) disagreed with the widespread stereotype that women speak more than men.
Moreover, with the aim of investigating whether there was strong consistency in the responses, the data were statistically analysed and presented in Table 4. The highest mean (M = 2.66, SD = .585) was found in the item "Men talk about sport more", with 50 participants (72.5%) agreeing with the statement. The lowest mean (M = 1.94, SD = .838), on the other hand, was calculated for the item "Men dominate conversation", where 26 participants (37.7%) disagreed and 22 of them (31.9%) agreed with the statement. The highest variability of distribution (SD = .889) was found in the responses to the assertion that men give more commands (only less than 7% of the male participants disagreed with the statement compared to around 44% of the female participants).
The lack of major consistencies was evident in this area as well. The data demonstrated high variance in responses, with only two items having a majority of respondents agreeing with the statements: the majority (n = 35, 50.7%) claimed that men swear more, and 50 respondents (72.5%) agreed that men talk about sport more. The highest percentage of agreement (72.5%) might be due to the widely accepted belief that men are generally more interested in sports and therefrom comes the conclusion that they speak about it more frequently. In terms of the latter, as was the case with the social norm of politeness, society also expects women to swear less, which is thus reflected in the language use as well.
The questions analysed above were concerned with specifically female and specifically male language. In addition, more general questions were posed to the participants to explore whether they held beliefs about characteristics of language usage potentially associable with different genders and about the relationship between gender and language.
A frequency analysis showed that the highest mean (M = 2.59, SD = .648) was found in the responses to the statement that they notice the differences in the ways men and women speak, where a majority of participants (n = 47, 68.1%) agreed. On the other end, the lowest mean (M = 1.75, SD = .755) was found in the response to the claim that they should use the language appropriate for their gender, where 30 participants (43.5%) disagreed with the statement (Table 5).

Do male and female participants differ in their perceptions of the way women and men speak?
In order to determine whether female and male participants in this study differed in their perceptions of the way women and men speak, an independentsamples t-test was conducted. The results showed no significant difference in the participants' perceptions of the way women speak (p = .415) between males (M = 2.20, SD = .426) and females (M = 2.30, SD = .413). Likewise, the results also showed no significant  Men interrupt more.  In order to get a more detailed insight into their answers, an analysis of individual items was required and thus conducted. An independent-samples t-test showed that most of the items (15 out of 16) for both the way women's and men's language use was perceived did not differ between female and male participants, which is numerically represented in p values in the table below. Only one item, the statement that men dominate conversation, demonstrated a significant difference (p = 0.40), with the frequency analysis showing that men were more likely to agree with the statement (46%) than women (28%). In addition to that, this analysis provided more details on the mean scores of individual responses, and the highest mean here (M = 2.97, SD = .843) was found in female participants' responses to the statement that men interrupt more, where 21 female participants (38.8%) agreed with the statement, while the lowest mean (M = 1.83, SD = .841) was found also among female participants, this time in response to the claim that men dominate conversation, where 24 (44,4%) of them disagreed. The data are shown in Table 7 below.

Do male and female participants differ in their perceptions of the relationship between language and gender?
An independent-samples t-test was run to determine whether female and male participants differ in their beliefs about gender-based differences. The results display no significant difference (p = .433) between the male (M = 2.32, SD = .368) and the female participants (M = 2.21, SD = .496) in the perceptions about the relationship between language and gender.
Further analysis showed that the difference in males' and females' perceptions about the relationship between language and gender for all individual items is insignificant, which is numerically represented in p values in Table 8. It can be noticed that the male participants obtained a slightly, but insignificantly higher mean overall and for individual items apart from the item stating "We should use the language appropriate for our gender", for which the female participants (M = 1.88, SD = .861) obtained a slightly higher mean than the male participants (M = 1.80, SD = .861).

Discussion
This study aimed to explore the perceptions of gender differences in language among students at International Burch University as well as to identify whether male and female participants significantly differ in these perceptions. Generally speaking, participants in this study do not seem to conform fully to some of the widely held stereotypes about male and female speech and about the relationship between language and gender, which supports our first hypothesis. Specifically, in terms of general beliefs about the relationship between language and gender, it is noteworthy that of 47 out of 69 of the study's participants (68.1%) claimed that they do notice the differences in the way men and women speak, but they do not believe as strongly that men and women are taught what language is appropriate for them, nor do they feel that they should speak according to their gender. It seems, therefore, that the participants believe in certain innate, perhaps biological differences women and men actualize in their language use, which are not necessarily taught by society. This question, however, would require a further, perhaps qualitative exploration of these beliefs in order to make any firm conclusions about them. That students are aware of the differences is in line with Kiełkiewicz-Janowiak and Pawelczyk's (2006) finding that people appear to believe in gender-based speech differences, even though they might not have witnessed them in reality.
Among the responses reflecting the stereotypes of the way women speak, two items showed major consistencies with over 50% of the participants agreeing with the statement -the claim that women are more polite yielded 56.5% (n = 39) agreement, while the one that women ask more questions was claimed by 59.4% (n = 41) of the participants. Both might be explained in terms of the social context in Bosnia and Herzegovina where women are expected to behave and thus talk more politely and are usually deemed more inquisitive and curious. Moreover, the latter could also be related to the negative belief of women being nosy and likely to gossip, thus their likelihood of asking questions is higher. Regarding other scientific findings, the results related to female politeness are in line with Coates' (2013) proposal of the stereotype still being widespread among general public as well as the results of Kramer's (1977) study in which this trait was associated with female speech. The findings of two studies of actual language use conducted by Brown (1998) and Holmes (1995) could be used here to draw a parallel between the stereotype and its footing in reality; namely, they both found that women do use more polite forms of language than men. In terms of the participants' agreement with the stereotype that women ask more questions, it reflects the findings of Pawelczyk's (2003) study in which respondents judged questions given for judgement as more likely to be produced by a female speaker as well as the study of gender stereotypes in Poland conducted by Kiełkiewicz-Janowiak and Pawelczyk (2006) in which women were believed to ask more questions "because they are nosy" (p. 360). However, it is important to emphasize that the study was conducted in a department in which most of the students are female, thus the participants might have gained the impression of women asking more questions, regardless of the analogous stereotype. On the other hand, the lowest mean (M = 2.05, SD = .855) was found in responses to the claim that women talk too much, which is in stark contrast to the results of some previous studies in which participants usually associated verbosity with female speech (Komarovsky, 1979;Kramer, 1974;Kramer, 1977). This is also quite an interesting finding since it might be used demonstrate that the younger, educated generation in the country is no longer as convinced in arguably one of the strongest stereotypes held by older generations.
In terms of the stereotypes of male speech, two major consistencies were found in agreement with the claims that men swear more (n = 35, 50.7%) and talk about sport more (n = 50, 72.5%). These stereotypes, among the others, perhaps most distinctly transcend the sphere of language, i.e. men are often perceived to be less cautious about their behavior in terms of politeness and are generally believed to like and participate in sports, which might be a plausible explanation for these results. Furthermore, they confirm some previous study's findings; for instance, the stereotypes of men swearing more was proposed by Coates (2013) and found in the studies of Kramer (1974Kramer ( , 1977 and Garcia-Zamor (1973). On the other hand, the lowest mean (M = 1.94, SD = .838) was found for the belief that men dominate conversation, which is quite an important finding taking into consideration that the notion of male dominance is a sort of a basis for any strong patriarchal narrative and environment. This result is in opposition to the findings of some other studies of stereotypes (e.g. Aries, 1996;Kramer, 1974Kramer, , 1977 as well as of real language use (e.g. Gilbert, 1990;Jenkins and Cheshire, 1990;O'Barr and Atkins, 1980). The second hypothesis stating that there is no statistically significant difference in the female and male students' perceptions of the way women and men speak was supported by this study. As for the individual items, only one significant difference was found (p = 0.04) for the claim that men dominate conversation, where more men (46%) than women (28%) agreed with the statement. This can be understood in the light of the male students' subjective perception of their own dominance in society in general, and consequently in conversational interactions as well. It seems, therefore, that the participants' gender generally did not have an effect on their attribution of stereotypes, which seems to be in congruence with findings from Kiełkiewicz-Janowiak and Pawelczyk's (2006) study in which 5 (20%) out of 25 stereotypes showed a significant difference between male and female participants (p. 362). Thus, the rejection of the hypothesis might be understood as an indicator that, if existent, certain stereotypes are spread equally across gender groups and that women and men do not perceive their gender's speech notably favourably or unfavourably.
As for the third hypothesis, stating that female and male students do not significantly differ in their perceptions of the relationship between language and gender, it was also supported -the t-test showed no significant differences among female and male responses. Namely, even though the students seem to be aware and somewhat conform to the beliefs, none of the items showed strong agreement within any gender group. This, however, seems quite expectable considering the fact that most participants were of the same nationality, Bosnian, and were thus raised in a culture which could perhaps be deemed homogeneous in terms of the norms it imposes and the beliefs it nurtures about gender. In addition, none of the statements in this category were positively biased towards any of the genders, which would have potentially yielded different responses along the gender variable. However, it should be noted that the male participants obtained a slightly, although insignificantly higher mean overall and for individual items, apart from the item stating "We should use the language appropriate for our gender", for which the female participants (M = 1.88, SD = .861) obtained a slightly higher mean than the male participants (M = 1.80, SD = .861).
Finally, it is important to note that the scope of the study is limited in different aspects, the most important one being the sample size. A larger sample would enrich the study population and therefore make the evidence more generalizable. Moreover, a more diverse sample, encompassing students from different universities and different study fields would also contribute to the significance of the paper. Since the participants in the present study were students at the Department of English Language and Literature at a private university in which the language of instruction is English, they are, due to the nature of their studies, certainly more aware of language and linguistic differences , as well as certain stereotypes pertaining to language and gender.

Conclusion
This study has led to a discovery that even though stereotypes seem to exist among university-level students in Sarajevo, Bosnia and Herzegovina, they do not seem to be as strong as to claim that young people in the country perceive women and men as strikingly different speakers. Considering the Bosnian context, having been extensively patriarchal and under the yoke of gender-based norms throughout its history, these findings seem rather surprising and could lead to a conclusion that the newer generations have experienced an evolution in thought through their education and exposure to more progressive, gendersensitive and equality-oriented ideas of the modern world, and that the progress has embraced both males and females. Additionally, this is a positive indicator that the entire society might be going through an important historical shift from a less patriarchal to a more egalitarian society. This discovery is important in as much as it provides a clearer image of gender stereotypes among the youth of Bosnia and Herzegovina, thus questioning the premise of the Bosnian society still being patriarchal in all its spheres -the belief which itself could be harmful because it potentially perpetuates the state in conviction that a significant change is impossible.
The discovery, however, could be further explored with focus on some important notions. Firstly, as proposed by Siegler & Siegler (1967), the stereotypes negatively portraying females' use of language might seem to have ceased to exist, but in reality, they have only become undesirable to express. Thus, another study should involve a more discrete strategy of uncovering the underlying beliefs through which subjects would unconsciously reveal what they actually hold to be true. Secondly, factors such as level of education, occupation, living area, etc., if considered, could portray a more detailed and thus more truthful representation of whether young people in the country have generally ceased to succumb to the overwhelming mentality or it is a matter of a wider sociocultural environment in which they live.