gender

Introduction

The U.S. is a concoction of many distinct cultures and languages as a result of settlement by non-natives and international students (Wharton University of Pennsylvania, 2017). There are also a variety of different English accents found in the U.S as a result of the people all over the world that are learning English. As of fall 2018, the percentage of public school students in the United States who were English as a second language speakers was 10.2 percent or 5.0 million students (National Center for Education Statistics). However, this does not mean that the U.S is free from prejudice and bias.

Sometimes accented speeches are harder to process and understand. When people are perceiving accented speech, they encounter difficulties that will reduce the “processing fluency”, as reported by Lev-Ari et al. (2010). Yet instead of simply perceiving the accented speech as harder to understand, people tend to perceive it as less truthful. One of the major groups of speakers affected by this is Korean speakers who grew up in a Korean-speaking environment. They tend to speak Korean-accented English due to not having been as exposed to standardized English (Linguistic Society of America). And so, the ESL Korean speakers could be perceived as less credible and intelligent, simply because they struggle with pronouncing some sounds in English and causing native speakers to have a harder time processing their speech.

The present study seeks to better understand how Korean ESL speakers who have Korean-accented pronunciation may be treated differently than standard American English speakers. This study explores whether listeners will perceive Korean accented English speakers as less favorable when compared to standard American accented English speakers.

Background

In Rickford & King’s (2016) study, a leading prosecution witness who spoke in African American Vernacular Accent (AAVE) had her crucial testimony dismissed. What led to this is that her speech was deemed incomprehensible and not credible. Unfortunately, this is not an unfamiliar experience for many people, including ESL Korean speakers. There are more than 81,000 Korean students studying in the United States, which accounts for 7% of the US’s international students population as stated by the United States Homeland Security (2015). Any non-native English speaker could face prejudice and unfair experiences in our society due to certain linguistic features that ESL speakers have (Lippi-Green, 2012). 

How was this study designed?

This research project focused on two parts: how non-English speakers are perceived socially about their intelligence level and credibility, and the perception of English speakers on Korean ESLs. The research was done by Google survey, and since our group members are all UCLA students, the participants were recruited on the campus. The survey was in two parts. The first part of the survey is generally asking the participants’ demographics and any experiences of racial discrimination. In the second part of the survey, the first thing that we did was ask their pre-existing bias about specific jobs’ credibilities and intelligence levels. To collect these data, we used the Likert scale to rate how they feel about teachers, lawyers, and doctors generally about their credibility/ intelligence level should be. Then we asked the participants to listen to three different recordings of reading English scripts by three different accents: Native English speaker, Korean ESL speaker who has a lot of Korean accent in it, and the third speaker who is fluent in English, has an American-ish accent, but have trouble pronouncing certain words. After listening to each recording, they were asked to rate how likely they would hire each recording’s voice as their teacher, lawyer, or a doctor. We selected these job fields because teachers, lawyers, and doctors are known to have high-intelligent, high credibility. The ratings were done on an interval scale of 1 (less likely) to 5 (most likely), and the ratings were done only after they listened to each recording.

Results & Analysis

From Table 1 and 2, we can see that most participants agree that the three professions—lawyers, teachers, and doctors— intelligence level and credibility lie between 4 or 5 on the Likert scale, which indicates that they agree the intelligence level and credibility of these professions are relatively high.

The data were run with two major systems: ANOVA and Post Hoc test. ANOVA gave an idea of whether there is a significant relationship between all three speakers, whereas the Post Hoc test compared each individual relationship between the three speakers. The following tables demonstrated the results of the ANOVA test as well as the Post Hoc test in terms of intelligence level and credibility. Note that speaker 1 represents the native speaker, speaker 2 represents the Korean accented speaker, and speaker 3 indicates the 3rd speaker.

The means of intelligence level of the three speakers are presented in Table 1. The native speaker has a mean of M = 3.01, the Korean accented speaker has a mean of M = 3.30, and the 3rd speaker has a mean of M = 3.55.

According to Table 1, the ANOVA test in terms of intelligence level demonstrated a p-value of less than o.oo1. It is known that a p-value less than 0.05 rejects the null hypothesis and indicates that it is statistically significant. Since we found statistically significant results by computing the ANOVA, we also ran a Post Hoc test to determine where the differences come from. The Post Hoc test suggested the mean differences of intelligence level between a native speaker and a Korean-accented speaker being 0.293; between a native speaker and 3 being 0.543; and between Korean-accented speaker and 3 being 0.250. The largest difference in means occurred between a native speaker and 3rd speaker with a native speaker mean of M=​​3.01 and a 3rd speaker mean of M=3.55 words. Although we found a significant relationship among the three speakers, the data does not support our hypothesis as we hypothesized that the Korean accented speaker’s intelligence level will be rated as the lowest, but the results suggest that the native speaker is rated as the lowest.

The means of credibility of the three speakers are presented in Table 3. The native speaker has a mean of M = 3.27, the Korean accented speaker has a mean of M = 3.25, and the 3rd speaker has a mean of M = 3.34.

Although in terms of credibility, the Korean accented speaker has the lowest rating compared to others, the ANOVA test in Table 2 demonstrated a p-value of p > o.05, which indicates that there are no significant differences. This result shows that the rate of credibility among the three speakers is similar, and thus reject our hypothesis that the Korean accented speaker will be perceived as less credible.

The results contradict our hypothesis. Our results suggest that Korean accented speakers are not perceived as less favorable in terms of both intelligence level and credibility.

Discussion

This research aimed to investigate how non-English speaking Koreans perceive socially and how others’ perspectives were. The ratings of credibilities and intelligence level were not as expected, and we are assuming that is due to the slower speaking pace that the native speaker recorded in the recording. Moreover, the ratings of Korean ESL speakers were not as expected and disconfirmed our hypothesis, which was that Korean ESL speakers were often treated less credible and less intelligent. However, there was a significant main effect on the intelligence level ratings between the three speakers. The third speaker with an American-ish accent was rated highest among other speakers, suggesting a third variable of speaking pace since the third speaker spoke at a fast speaking speed.

There are some more limitations in the study, such as gender bias—we only used voice recordings of male speakers. So we might want to include female recordings to avoid gender bias. Secondly, there might be a participant bias, meaning participants might assume what our study was about and rate accordingly. We also see some tendencies in the job fields; for instance, teachers were not rated as expected in the pre-rating survey. They were rated not highly enough than our expectation on both credibility and intelligence level. Since the recording was quite long, there might be selective attrition. Participants might feel bored and drop out during the experiment and rate every Likert scale as 3, which is neutral. Lastly, the sample size used was very small. Most participants were UCLA liberal arts students, and most of them were our friends who are also Korean ESL speakers. Since the participants are our friends, they could recognize our voice recordings during the experiment, which could affect the ratings. Moreover, since the sample size was small and limited, the study cannot generalize all Korean ESL students. We hope further research will consider speaking pace for future directions—for example, the relation of speaking speed and one’s perception. Or removing the slow speaking rate and making all the recording speakers talk about the same speaking pace. Moreover, to investigate further, we suggest using a Latin square counterbalanced 2×2 factorial design to see the interaction between gender and accented speech in the perception of credibility and intelligence level. So the first Independent variable would be gender (male, female), and the second Independent variable would be the accent with two levels: accented speech and non-accented speech. Further research would need more diverse participants in age and demographic. And hoping the further study could research other East Asian ESL speakers, not only limited to Korean ESL speakers.

References

English Language Learners in Public Schools. National Center for Education Statistics. (n.d.). Retrieved November 12, 2021, from https://nces.ed.gov/programs/coe/indicator/cgf.

FAQ: Why Do Some People Have an Accent? Linguistic Society of America. (n.d.).

Retrieved November 12, 2021, from https://www.linguisticsociety.org/resource/faq-why-do-some-people-have-accent.

Lev-Ari, S., & Keysar, B. (2010, June 25). Why don’t we believe non-native speakers? the influence of accent on credibility. Journal of Experimental Social Psychology. Retrieved November 12, 2021, from https://www.sciencedirect.com/science/article/pii/S0022103110001459

Lippi-Green, R. (2012). English with an accent: Language, ideology and discrimination in the United States. Routledge.

Rickford, J.R., & King, S. (2016). Language and linguistics on trial: Hearing Rachel Jeantel  (and other vernacular speakers) in the courtroom and beyond. Language 92(4), 948-988. Doi: 10.1353/lan.2016.0078.

‘speaking American’: Regions, accents and the subtleties of language. Knowledge@Wharton. (2017, January 18). Retrieved November 12, 2021, from https://knowledge.wharton.upenn.edu/article/speaking-american-regions-accents-and-thesubtletis-of-language/.

Streamers in our research

This experiment revolved around three male streamers and three female streamers. The three male streamers who were analyzed were: AverageJonas, itsRyanHiga, and Ethos and the three female streamers are Kyedae, Ploo, Tiffae. These six streamers were chosen due to their popularity/recognition, skills, and various personalities.

 The male streamers are around 30 years old, and the females are around 20. They are mostly all of white or East Asian descent and are English speakers. To be more specific, AverageJonas is a 32-year-old Norwegian English-speaking streamer, itsRyanHiga is a 31-year-old Japanese American English-speaking streamer, and Ethos is of unknown age and is an Asian American English speaking streamer. Kyedae is a 19-year-old Japanese Canadian English-speaking streamer, Ploo is a 22-year-old Asian American English-speaking streamer, and Tiffae is a 27-year-old Taiwanese English-speaking streamer.

What we already know

Based on gender stereotypes, men are perceived to be louder, more aggressive, and curse and insult more, and also since gaming is seen as a “male-dominant” space, we also thought gaming terms would be more male-gendered as well. Because of that, we decided to analyze vocabulary reactions to certain game events by gender.

We also analyzed the vocabulary differences by comparing streamers in an all-male team and a mixed-gender team. Previous research states that men in an all-male friend group played into male stereotypes in their group interactions more than those who had their girlfriend or female friend in their groups (Cameron 1997). In our experiment, we thought that these previous findings could suggest that female streamers may also feel pressure to conform to their male teammates and speak in a more stereotypically male manner, so that was an aspect we decided to look at.

We decided to analyze streamer and fan interactions between genders. Previous research stated that women communicate to build relationships, so we thought we would see that when observing how streamers interact with their viewer’s donations (Van Herk, 2012).

Methods

For the experiment, we watched six streamers for key vocabulary words in their speech and their chat interactions. The streamer’s vocabulary choices were analyzed when they reacted to these three events: donations from fans, winning a game, and losing a game. The reactions to those events were grouped by event and gender- for example, one group would be “Vocabulary in reaction to winning a game from a male streamer.” Then the groups would be compared across the same event and against opposite genders. An example comparison would be comparing reactions to winning by male streamers against responses to winning by female streamers. We watched each streamer stream for an average of 2 to 3 hours in total.

While we were watching the stream, we noted down some vocabulary and phrases used by the streamer. Then we sorted phrases between each scenario and compared the differences. After we sorted the words and phrases, we looked at the difference between streamers and gender. We also decided to analyze our overall results of vocabulary choice based on the gender of the streamer’s teammates.

Thesis

We found a clearer divide in the vocabulary when we compared streamers that played with all-male team and streamers that played in a mixed-gender team. Through our results, we can conclude that while there were differences in a streamer’s vocabulary due to the streamer’s gender, the main differences relied more on the gender of their teammates. Streamers with all-male teammates used more gaming-specific terms, insults, and cursed more. But as for the streamer’s own gender, we found that females were more likely to negatively refer to themselves in the face of a negative event, while males would blame the game and enemy team. We also found that females were more likely to initiate and maintain friendly interactions with their viewers and team.

Results / Analysis

Win reactions based on streamer’s own gender

To start off, we compared the vocabulary that females used when they won a game versus the vocabulary used by males when they won a game.

Female wins

As you can see from the word cloud generated from the data across the three female streamers, we identified that female streamers didn’t use any swear words in the situation of a win. Female streamers tended to comment on the win as the group’s achievement, using phrases that begin with the subject “we”. Two out of the three streamers used different variations of the phrase “oh my god” which is commonly associated with the regional marker of valley girls. Male wins

Looking at the result of the male streamers, we saw that male streamers would swear after a win. One of the male streamers, Ryan Higa used the Sh-word after a win. Overall, when commenting on a win, the male streamers tend to phrase the comment as talking about the game and not how well they played as a group(we).

Through these word clouds, we can see that there’s gendered language – the females kept their female gender identity through speech reminiscent of valley girls, despite being in a male-dominated space, and males, true to our hypothesis, used more crude language.

Loss reactions based on streamer’s own gender

Next, we analyzed the reactions of losses by females and by males.

Female loss

The word cloud on the right shows the utterances of the female streamers when facing the scenario of a loss. Unlike the result from utterances after wins, we saw swear words. The F-word was used in the phrase “oh my f**king god”. Another commonality noticed between the female streamers is that they tend to address the loss of the team as their responsibility, using the subject “I” when commenting on the loss, for example, “I messed up” or “I can’t win a game”.  Female streamers also used encouraging phrases such as “that was a very nice try” to show support towards teammates.

Male loss

The male players also used swear words after a loss, for example, the F-word was used in the phrases. In general, they form their phrases without criticizing the loss as their responsibility or their team’s responsibility. Instead, they use phrases that begin with “that”, “they” or “this girl” to address the game or the other team/team member. 

The main difference between female reactions to losses and male reactions was not what we expected- we saw both genders swear (but males to a greater degree). The biggest difference was actually the target of negative or aggressive language. The females tended to blame themselves for the loss and put themselves as the target of any negative descriptors or words. Males would blame the game or the enemy team. We also saw that females, to further put the blame solely on themselves, would congratulate their teammates.

Donations / subscriptions interactions based on streamer’s own gender

We then looked at reactions from females and males to donations and subscriptions by their fans. We mainly looked at the manner in which the streamers interacted with their fans, and we found that all-male streamers used the same phrase for all their donations and subscriptions: Thank you XXX (viewer’s name) for the donation/subscription. Two of the female streamers also used this format; however, one of the female streamers interacted with the audience by asking about their life and would reciprocate the viewer telling her “Peace and love” by saying “Peace and love” back. This indicates that female streamers demonstrate a rapport style and make an effort to build relationships in their communicative behavior. This agrees with Deborah Tannen’s claims, suggesting that women are more likely to use language to build and maintain relationships while men are more likely to use language to communicate factual information (Van Herk, 2012).

Overall vocabulary results based on teammate’s gender:

While we conducted our research, we noticed that there seemed to be a divide in gendered terms not by the gender of the streamer, but by the gender of the teammates that the streamer played with. We aggregated our vocabulary results and split them by the gender of their teammates, resulting in the bar chart below.

Streamers playing with all-male gamers said many more swear words (f*** words, sh** in particular) and gamer terms (frags, bait, camp, pick, etc), while streamers playing with a mixed-gender team did not say any swear words from our analysis and fewer gamer terms. This is due to in particular how male streamers feel the need to act more masculine and also use more common lexical entries when with other male teammates only. However, when female streamers were with female gamers or mixed gendered teammates, they tend to overall be calmer and use more relaxed language.

Another difference that was noticed was the patterns that male streamers and female streamers exhibit during the same scenarios were unexpected. For example, when male streamers and female streamers make the call-out that the enemies were attacking the B site, Kyedae says “They’re going B, right?” and Ethos says “They’re going B, they’re going B.” Kyedae said this scenario to a teammate in more of a questioning, unsure manner even though the enemies were clearly rushing B site. Ethos, however, said his callouts in a confident, command-like manner. Furthermore, what was interesting was that even if the call-out was wrong, female streamers tend to speak in a safer question-like manner, but male streamers spoke in confident manners then apologize when they were wrong.

Conclusion

Since we analyzed male and female speech in a heavily male-dominant space, we believe that our findings could apply to other male-dominant spaces. We looked at references to a similarly male-dominated industry- the sports entertainment industry. In Eastman & Billings (2000), they found that the tone used by commentators when describing male athletic wins was “enthusiastic…but derogatory” towards women’s athletic wins. This seems to relate to our findings that female streamers were more negative towards themselves and could suggest a larger connection to how derogatory attitudes towards women’s achievements in male-dominated spaces prevails and can even permeate to the women themselves. Also considering that E-Sports is still an emerging field, it’s plausible to say that the previous biases that existed in older male-dominated industries like the sports entertainment one examined in Eastman & Billings have been long established and continue to reappear today.

As for future directions, we looked mainly at the gender of streamers and controlled on popularity, the game played, and the scenarios reacted to. But through our study, we found that there were a lot more factors that could be confounding. For example, the players had varying skill levels that could have swayed our results, so we could better control for that. Another factor we saw was the gender of the teammates our streamers played with. We saw a stronger correlation between the gender of the teammates and gendered lexical entries. It would be an interesting follow-up study to analyze the interactions of an all-female team such as Cloud9 White, versus an all-male team’s interactions.

References

Çakır, G. (2021, March 5). Twitch slang and common terms explained. Dot Esports. Retrieved November 12, 2021, from https://dotesports.com/streaming/news/twitch-slang-and-common-terms-explained

Cameron, D. (1998). Performing gender identity. Language and gender: A reader.

Clayton, N. (2021, February 24). Valorant wants more women and ‘minority genders’ competing at its highest levels. pcgamer. Retrieved November 12, 2021, from https://www.pcgamer.com/valorant-wants-more-women-and-minority-genders-competing-at-its-highest-levels/

Eastman, S.T., & Billings, A.C. (2000). Sportscasting and sports reporting: The power of gender bias. Journal of Sport and Social Issues, 24, (2), 192-213

How many people play valorant? player count tracker (2021). Dexerto. (2021, October 4).         Retrieved November 12, 2021, from    https://www.dexerto.com/valorant/how-many-people-play-valorant-player-count-tracker-2021-1668158/.

Messner, M.A., Duncan, M.C., & Jenson, K. (1993). Separating the men from girls: The gendered language of televised sports. Gender and Society, 7 (1), 121-137

Nakandala, S. (2016, November 22). Gendered Conversation in a Social Game-Streaming Platform. Gender Conversation in a Social Game-Streaming Platform . Retrieved from https://www.researchgate.net/profile/Supun-Nakandala/publication/310611118_Gendered_Conversation_in_a_Social_Game-Streaming_Platform/links/5839c31808aef00f3bfbbbf1/Gendered-Conversation-in-a-Social-Game-Streaming-Platform.pdf.

Pellicone, A. J. (2017, May 6). The game of performing play: Understanding streaming as Cultural Production. Retrieved November 12, 2021, from http://library.usc.edu.ph/ACM/CHI%202017/1proc/p4863.pdf

Takahashi, D. (2021, June 14). How riot games will ensure that Valorant’s esports stars include women. VentureBeat. Retrieved November 12, 2021, from https://venturebeat.com/2021/06/14/how-riot-games-wants-to-ensure-that-valorants-esports-stars-include-women/

Van Herk, G. (2012). What is sociolinguistics? (Vol. 6). John Wiley & Sons.

In its everyday use, language is constantly changing. It adapts to best fit the current needs of a society and reflect what is culturally relevant. When analyzed correctly, language is a record of a culture – relating norms, cultural roles, dynamics, etc.

Take a look at the following quote:

“The father, what he wanted was that his daughter leave the house dressed in white and that was his goal” (Perez, excerpt 8).

Without further context, this excerpt allows for some insight as to what gender roles and familial expectations characterized this speaker’s community. Here, there is an emphasis on community and culture, because the speaker’s expectations have been determined by the interactions within their own community, which in turn, is reflected in their use of language.

This is not a new take on language and can be tied back to symbolic interactionism. This theory posits that meaning is constructed through everyday interactions and conversations (“Symbolic Interactionism Theory”, 2014). In other words, much of what we come to value, identify with, expect, draw meaning from, etc., was learned through our interactions with others within our community.

This is the framework that formed the basis of the present study. With a focus on word choice, our interest was with Latina immigrant mothers, women who grew up within Latin American norms and expectations, then found themselves having to immerse themselves into a new culture and unfamiliar social contexts (often resulting in conflicting interactions).

According to prior research, Latina immigrant mothers’ perceptions of education are impacted by interactions in both American and Latin America social contexts. To illustrate, in Mexican culture, women are often expected to prioritize familial obligations over their education (Perez, 2009). In addition, women in Mexico are expected to be most involved in their children’s lives, while in the U.S., it is common for both mother and father to work (resulting in comparatively less involvement) (Perez, 2009). Overall, prior findings support the assertion that immigrant women often form judgements from experiences and meaning created during early-life interactions (Perez, 2009).

With the intention of seeing how these findings manifested within an immigrant mother’s direct interactions with their children, we were looking to explore the following: What factors influence the expectations of immigrant Latina mothers? How do they communicate this to their children?

The study’s analysis was collected via informal interviews with two Latina immigrant mothers. Mother 1 is from Jalisco, Mexico. She immigrated to the U.S. 41 years ago, at age 8. Her two daughters are 19 and 16 years old. Mother 2 is from Oaxaca, Mexico. She immigrated to the U.S. 30 years ago, at age 16. She has two daughters, ages 19 and 23.

We hypothesized that Latina immigrant mothers have higher expectations for their older daughters. These expectations take the form of expecting the daughters to attain a higher education. We further hypothesized that Latina immigrant mothers would demonstrate these ideals through word choice, specifically, through the use of more positively correlated words and less fillers when being interviewed about their daughters’ educational goals.

In order to investigate our hypothesis, we had each daughter interview their respective mothers- mother 1 was interviewed by her oldest and youngest daughter and mother 2 was interviewed by her oldest and youngest daughter. The younger and older daughters each asked their mother the same series of ten questions (see the list of questions below), but they interviewed their mothers on different days and asked the questions in different orders. Each of the younger daughters asked the interview questions on the first day, and, then, a few days later the older daughters interviewed their mothers.

We recorded these interviews in order to be able to transcribe them so that we could make note of the following nuances in the mothers’ speech patterns: pauses, filler words, connotative word choice, levels of detail and any phrases that may allude back to gendered expectations. For the purpose of our research, we took pauses longer than two seconds into consideration; we did not analyze one second or less than one second pauses because we interpreted them as the mothers taking a moment to formulate their thoughts. We analyzed pauses longer than two seconds as an indicator of hesitancy on the mothers’ behalf about how they should respond to the question that was posed to them. We counted all the instances of filler words such as “Ah,” “Ahm,” “Uh,” “Uhm,” “Eh,” “Ehm,” and “Mmm ”as markers of hesitancy when preceded or followed by a pause as well.

We also used the context of the mothers speech to determine what pauses or fillers the mothers used to give themselves time to think and what pauses or fillers the mothers used when hesitant about how to deliver their responses. In addition to that, we also focused on words such as “orgullosa,” meaning proud,” or “gusta,” meaning “like,” as samples of connotative word choice in order to quantify how approving or disapproving the mother was in her response. We noted the amount of detail mothers went into in their responses for each question from the interview with their youngest daughter and cross referenced it with the length of their response from the interview with their oldest daughter. We made note of this, in order to determine if going into more or less detail with one daughter or the other had any implications in regards to their expectations for each daughter. While doing all of that, we also analyzed the data for any instances of references to gendered expectations the mothers brought up to draw conclusions about how well the educational expectations the mothers held for their daughters aligned with gendered expectations that are typical of Latinx culture.

Before we discuss our results, click this link: Latina Mothers in order to hear the audio.

The word choices in the recorded interactions were meant to indicate hesitancy or show intention in the responses, with a special focus on the answers from the mother. Filler words and phrases were tallied specifically with consideration that it might mean there was more difficulty finding a word, alluding that mothers felt the need to be more intentional with their responses. While there was a very large contrast between the amount of them the mothers used, the results indicated that amongst them both, they used more filler words with their older daughters. Specifically Mother 2 used 17 filler words with her youngest daughter and 23 with her oldest daughter. Similarly, Mother 1 used 2 filler words with her youngest daughter and 4 with her older daughter. This observation was especially interesting, given that in both cases the older daughters interviewed their mothers after the younger ones had so despite having already been primed 2 days prior, there was still more use of filler words with older daughters.

Pauses had the opposite results. When tallying we counted pauses above two seconds, and found that both mothers used more pauses when communicating with their younger daughters. In this case the mothers used a similar amount of pauses despite the location varying. Mother 2 used 16 pauses with the younger daughter, compared to just 13 pauses with the older one. Mother 1 used 13 pauses with the younger daughter and 12 pauses with the older daughter. Analyzing pauses was done with the intention of considering their hesitation when answering questions in a certain way. Overall this analysis indicated that when speaking with their younger daughters, the mothers used more pauses for their younger daughters. This could signify that the word choice required more thought and was more selective. However, it is also important to recognize that this length in pauses could be attributed to the younger daughters interviewing their mother first.

The next part of the analysis consisted of finding words that held strong positive or negative connotations. We immediately noticed that there were not any explicit words that had negative meaning or were used to allude to something in a negative light. On the other hand, we were pleasantly surprised with the amount of words with positive meaning. Although they were used in varying contexts, both mothers expressed about the same amount of positive words. Most of these words overlapped in all conversations, with some of the most recurrent being “orgullosa” meaning ‘proud’, “exitosa” meaning ‘successful’, and “gusta” meaning ‘liking/liked’. Each word was interpreted as approval of their accomplishments or what they’re doing academically. These positive words were evenly spread among all interviews and didn’t indicate any biases for stricter expectations with one or the other daughters. One interaction that exemplified the role of maternal figures in setting expectations, was Mother 1’s response to the question “What do you think about my goals compared to the cultural expectations you grew up with?”. She responded with, “Esta buena pregunta para tus abuelas”, ‘this is a good question for your grandmothers’ possibly implying that expectations for women are set by the women that raised them.

Finally, our study analyzed the level of detail provided in the responses. The purpose of taking detail into consideration was to see if the mother was involved more with one daughter than the other or if the mother felt the need to explain her responses because of a more demanding expectation to either of her daughters. While the answers varied in length, there was an increased length in the responses for the older daughters from both mothers. This again could indicate that mothers feel the need to reinforce specific goals and expectations more thoroughly to their older daughters.

However, in studying this we observed other responses that supported our hypothesis that Latina immigrant mothers held higher expectations for their elder daughters in comparison to their younger daughters. The first question being “How do you feel about my career choices?” where Mother 1 explains to both daughters that she is content in their choices. However, she then goes on to note that her older daughter’s aspirations are rather competitive and she wants to see them all the way through. This was in contrast with her response to her younger daughter where she suggested a career in Broadway, which is much more difficult to land a secure spot in. This response perhaps shows that the mother wants her older daughter to be more secure in her career choice, followed by the expectations of being in a stable job — expectations not found in the interview with her younger daughter.

In summary, our study found that with the cases given, we were able to find some indications of older daughters being held to a higher career standard. Such instances were found in word choices analyses regarding careers and education. Specifically, we were able to find patterns of correlating pauses and filler words, and were even able to find patterns of expressing more detailed answers to older daughters. However, due to certain limitations (to be specified in subsequent paragraphs), it is clear that these findings are not substantial enough for a definitive conclusion. 

Our research would benefit from further research to argue for a more solid conclusion. Additional exploration and analysis on our test subjects and methods could provide more substantial evidence for whether or not Latina immigrant mothers have higher expectations for their first eldest daughters compared to the youngest.

Our research limitations consisted of a few things. First, out of the interest of time, our interviewees were only allowed ten questions to ask their mothers. These questions did not include follow up questions, which would have greatly expanded on certain ideas. For example, one instance that could have benefitted from more explanations from both responses is to the question “How involved do you think you were and are in my school life and extra curricular activities?”.

To expand, one mother replied with “..un ~ochenta y cinco~ por ciento posiblemente” ‘possibly 85 percent’  for both daughters. This mother only expands briefly when talking to her older daughter, ranging her involvement between elementary school and high school. However, one way to have gained more qualitative data would have been to ask specifically in what ways this mother came to the conclusion of 80 percent.

The other mother responded with “I am very involved in the activities and studies of my daughters”  for the younger daughter and “Very, very much in school” for the older daughter. Similar to the previous mother, a follow up question of asking “How so?” or “In what ways?” could have provided more data to compare to with the first mother, in regards to what these two mothers did to stay involved.

Moreover, our study only used two mothers, therefore, our work should be considered as a case study rather than extensive research. It is only a small representation of the Latina motherhood community. On a similar note, with only two mothers, it was difficult to compare and draw conclusive statements. However, if there were more resources, subjects, and time allowed, these limitations could be reduced.

These findings may contribute to a larger phenomenon on who maintains such expectations in a Latina community, as well as connecting to as far as how this affects a woman’s social identity in her aspirations. 

References

Aviles, G. (2019, November 20). More Latinas have college degrees, but they’re still making less than white males. NBCNews.com. https://www.nbcnews.com/news/latino/more-latinas-have-college-degrees-they-re-still-making-three-n1087101

Clark, H. H., & Fox Tree, J. E. (2002, February 27). Using uh and um in spontaneous speaking. Retrieved from http://www.columbia.edu/~rmk7/HC/HC_Readings/Clark_Fox.pdf

Crosnoe, R., Ansari, A., Purtell, K. M., & Wu, N. (2016). Latin American Immigration, Maternal Education, and Approaches to Managing Children’s Schooling in the United States. Journal of marriage and the family, 78(1), 60–74. https://doi.org/10.1111/jomf.12250

Field research. (n.d.). Retrieved May 11, 2021, from https://www.researchconnections.org/childcare/datamethods/fieldresearch.jsp

Mendoza-Denton, N. (2008) “THAT’S THE WHOLE THING [tɪŋ]!”: DISCOURSE MARKERS AND TEENAGE SPEECH. In J. McIntosh (Ed.), Homegirls (pp. 265-291). Wiley.

Perez, A. (2009). Understanding Attitudes about Education within two Cultural Contexts: Comparing the Perceptions and Expectations of Mexican Mothers in the Greater Los Angeles Area to their Cohort in Mexico. Graduate Studies At California Lutheran  University.

Symbolic Interactionism Theory. Communication Theory. (2014). Retrieved May 2021, from https://www.communicationtheory.org/symbolic-interactionism-theory/.

For our study, we chose a convenience sample of three participants. The target demographic were nonbinary English-speakers in the Millennial/Gen Z age-range. One participant was assigned-female-at-birth and part of the cisgender control group. The other two participants were agender, and this group was a mix of assigned-female-at-birth and assigned-male-at-birth. Each participant provided at least six formal email threads for analysis. At least three of these were addressed to male recipients while the other half were addressed to female recipients. All recipients held a position of power over the participants. Before examining the emails, we produced a codebook full of key words and markers which signify powerless language. Here are a few examples of markers we coded:

We tallied these codes to calculate the average frequency of powerless markers for every X amount of words in each message.

Gerard Van Herk’s What is Sociolinguistics? (2017) provides concepts of gender identity, performance, and practice. Herk asserts that “in cross-gender interactions we can expect to see the greatest gender effect,” therefore, we analyzed cross-gender interactions to help gauge the effect of perceived gender.

All data were coded twice, once by each of the researchers who did not compose said text. After coding was completed, we compared our findings in order to determine percent agreement of coding and in order to confirm that our coding system is well-defined and useful. Below are some example emails which we coded and analyzed as part of our data pool. The first is written by our assigned-female cisgender participant addressed to a female recipient, and the second is written by our assigned-male agender participant addressed to a male recipient. All names and sensitive information have been anonymized.

We found in our results that there is a notable increase in powerless markers in cross-gender interactions, according to assigned gender at birth. Below is a table summarizing the average frequency of powerless language towards female and male recipients for each participant:

We can see that gender identity had little, if any, effect on speech in emails. Regardless of identity, cross-gender interactions according to assigned gender at birth still produced more exaggerated use of powerless language, just as Van Herk predicted. As researchers, we expected that our agender participants, Amanda and Erik, would not conform to binary patterns. We predicted that their average frequency of powerless markers per message would be roughly the same across differently gendered recipients. Instead, our actual results imply that binary socialization, upbringing, and external perception impact the way that nonbinary people navigate the world. These results have many practical applications for the LGBTQ+ community. Some nonbinary people may be okay with presenting in a way that aligns with their assigned gender at birth; however, others may desire to present a different image. Perhaps agender and other nonbinary individuals can use this research to consciously move towards less gendered speech, if they so desire. Furthermore, individuals wishing to present as more masculine or more feminine may also use this research to alter their speech so that it better aligns with their identity.

However, it is also important to note that these cross-gender disparities are potentially social constructs. Powerless language is not inherently feminine or masculine, but we associate these qualities with gender performativity.

If a person is in a position of less power, they are inclined to use more polite language. This is not necessarily powerless language in itself. In fact, it can be powerful language in its own right, as the speaker or writer implements manipulative tactics in request-making. For those in a lower position, polite language is more likely to produce results.

There are several interesting topics for further research. As mentioned above, upbringing may have a significant impact on cross-gender language. Patriarchal, matriarchal, and neutral households may produce differing attitudes towards the power dynamic of men and women. So, depending on a person’s upbringing, they may use either more or less powerless language when speaking to different genders.

The sociolinguistic study of gendered language is said to have started with Lakoff’s early studies and eventually his book “Language and Woman’s Place” (1975). Since then, the study of language and gender has developed greatly. Lakoff’s Research has been referenced and expanded upon many times, but generally focuses on a binary view of gender, and non-binary topics are under researched fields of study. Although both gender identities and social genders outside of the binary have always existed, the term ‘agender,’ “denoting or relating to a person who does not identify themselves [sic] as having a particular gender” (Google Definitions), was first coined on the internet in the year 2000 (Them). Our own research is extremely limited, examining a small pool of data from a convenience sample of three participants, only two of which are agender. We did not study other nonbinary identities outside of agender, nor did we study binary transgender people. Further research may reveal more patterns or nuances which escaped our own research.

References

Brown, P., & Levinson, S. C. (2007). Politeness: Some universals in language usage. In Politeness: Some universals in language usage (pp. 61-83). Cambridge: Cambridge Univ. Press.

Cameron, D. (2010). Performing Gender Identity Young Men’s Talk and the Construction of Heterosexual Masculinity. In 1367818629 1000231388 D. Kulick (Author), Language and sexuality. Cambridge: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511791178

Herk, V. G. (2018). What is sociolinguistics? In What is sociolinguistics? (pp. 86-100). Hoboken, NJ: John Wiley & Sons.

Myers, Kristen. (2012). Cowboy Up!. ResearchGate. www.researchgate.net/publication/276215982_Cowboy_Up.

O’Barr, William M., and Bowman K. Atkins. ‘Women’s Language’ or ‘Powerless Language’? Women and Language in Literature and Society, by Sally McConnel-Ginet (1980). New York: Praeger, pp. 93–110.

Silverstein, M. (1976). Shifters, Linguistic Categories, and Cultural Description. In 1367858396 1000257581 K. H. Basso & 1367858397 1000257581 H. A. Selby (Authors), Meaning in anthropology (pp. 11-55).

Them. (2018). What Does It Mean to Be Agender? Them., Them. www.them.us/story/inqueery-agender.