Gender-fair language strategies often face resistance. Negative attitudes have been documented against specific gender-fair reforms, such as the replacement of the masculine generic he with the paired form he/she (Blaubergs 1980), guidelines for non-sexist language (Parks and Roberton 1998), and the introduction of a gender-neutral pronoun (Gustafsson Sendén et al. 2015; Bäck et al. 2015, 2018; Lindqvist et al. 2019).

Past research has discerned several arguments that are used against adopting gender-fair language. Blaubergs (1980) developed a taxonomy of arguments against gender-fair language in the wake of the proposal in the 1970s to replace the masculine generic he with he or she. Blaubergs analyzed a sample of arguments in newspapers articles, scientific journals and other media, and established eight categories of arguments: Cross-Cultural, Language Is a Trivial Concern, Freedom of Speech/Unjustified Coercion, Sexist Language Is Not Sexist, Word Etymology, Appeal to Authority, Change Is Too Difficult, Inconvenient, Impractical or Whatever (hereafter shortened to Change Is Too Difficult), and It Would Destroy Historical Authenticity and Literary Works (hereafter shortened to Historical Authenticity). Parks and Roberton (1998) extended Blaubergs’ taxonomy based on undergraduate’s arguments against gender-fair language with four more categories: Sexism Is Acceptable, Hostility Toward Proponents of Change, Tradition, and Lack of Understanding.

In Swedish, the gender-neutral third-person pronoun singular hen has been introduced as a complement to she (hon) and he (han) and is used as both a pronoun to refer to individuals with non-binary gender identities and as a generic pronoun (SAOL 2014). Hen’s introduction received a lot of media attention in 2011, and in 2014 the pronoun was included in the Swedish dictionary. At first, there were heated debates in the media, in universities, and among laypeople (Milles 2013). In contrast to past reforms that added gendered (often feminine) forms to make women more salient, a gender-neutral pronoun reform reduces gender information. The present study examines whether arguments against the gender-neutral hen are similar to arguments from past gender-fair language reforms using feminization strategies in other cultural contexts. To this aim, the taxonomies of criticism developed by Blaubergs (1980) and Parks and Roberton (1998) were used to categorize the arguments against the gender-neutral pronoun hen.

Gender-Fair Language Planning

Feminist scholars have promoted gender-fair language for over half a century. In the 1970s, feminists considered the generic use of masculine pronouns and masculine occupational titles problematic and “both a symptom and a source of fundamental androcentrism” (Braun et al. 2005, p. 3). Empirical studies have shown that masculine generics are androcentric because they more readily evoke mental images of men (Gastil 1990; Moulton et al. 1978). This connotation was for example shown in German where the generic masculine forms of roles and occupations were associated with men more frequently than with women (Stahlberg et al. 2001).

Furthermore, masculine generics influence attitudes and behavior, especially among women. For example, women were less motivated to apply for a job when masculine generics were used in a job advertisement (Bem and Bem 1973). Women also felt less belonging, motivation, and identification with a job when masculine generic pronouns were used in a mock interview compared to gender-fair alternatives (e.g., he or she; one; Stout and Dasgupta 2011). Social judgments can also be influenced by masculine generics. Women applying for jobs were judged as less suitable for a high-status leadership position when masculine forms were used in comparison to paired forms (Horvath and Sczesny 2015).

To counter the negative effects of masculine generics, many languages have introduced gender-fair alternatives to masculine generics throughout the late twentieth century. These initiatives have focused on creating either paired forms that include references to both women and men (called feminization because the feminine form is added) or gender-neutral forms (called neutralization because gender information is reduced; Sczesny et al. 2016). The most common way of pairing gendered pronoun forms is “he or she” (Willis and Jozkowski 2017). Gendered pronouns can also be paired as “he/she,” “s/he,” “he (she),” or they can be alternated throughout a text. Critiques of paired pronoun forms include awkwardness in use, that constructions like “s/he” cannot be said out loud, and that there are no comparable grammatical case forms, such as “his or hers” (Madson and Hessling 1999). Even organizations that deter the generic use of “he,” such as the American Psychological Association, discourage the repeated use of paired forms stating that “the repetition can become tiresome” and that forms such as he/she or (s)he are “awkward and distracting” (American Psychological Association 2009, p. 74).

Paired forms also suffer from an androcentric effect called “male firstness” (Willis and Jozkowski 2017) because “he or she” is more common than “she or he.” Male firstness has also been documented for personal names (Hegarty et al. 2011) and in scientific articles that present gender differences (Hegarty and Buechel 2006; Willis and Jozkowski 2017). In addition, paired forms emphasize gender as a dichotomy by explicitly denoting gender as constituting the binary categories woman (she) or man (he), excluding individuals with non-binary gender identities (Ansara and Hegarty 2016; Hyde et al. 2018). Hen is used as both a pronoun to refer to individuals with non-binary gender identities or as generic form (SAOL 2014). In this way, gender-neutral pronouns are different from paired language reforms because they make non-binary identities visible in language and decrease the dichotomous perceptions of gender (Wayne 2005).

The Introduction of Hen

The introduction of hen generated a heated debate in 2011 (Bäck et al. 2015, 2018; Lindqvist et al. 2016; Gustafsson Sendén et al. 2015), despite Sweden being an egalitarian country with a history of adopting feminist values and language reforms (Milles 2011). Past studies have shown that egalitarianism is related to a more positive opinion of gender-fair language (Formanowicz et al. 2015) and a greater use of gender-fair language (Hodel et al. 2017). However, during the debate in Sweden, the response was mostly negative. In a study documenting the attitudes in a representative sample collected in 2014 in Sweden, 55% of the participants expressed negative attitudes toward hen whereas only 14% were positive (Bäck et al. 2018).

The present study focuses on the contents of arguments that Swedish people use against gender-neutral pronouns. To this aim, we use the categories of arguments against gender-fair language established by Blaubergs (1980) and Parks and Roberton (1998) to investigate whether their taxonomies are also valid for gender-fair language reforms that introduce gender-neutral pronouns (neutralization in contrast to feminization) in a different time and cultural context. If the arguments fit past taxonomies, this continuity might indicate that people are more averse to gender-fair language change than to the content of the reforms themselves.

Method

Participants and Procedure

In the present study, arguments against hen were the focus. Therefore, the sample included participants who were familiar with hen and expressed at least one argument against using hen. First, participants read a definition of hen: “Hen could be used when it is not necessary to specify a gender, for example to replace ‘he/she,’ or for individuals that don’t want to categorize themselves as she or he.” Subsequently, they described in a free-text response why they did or did not want to use hen. Participants (n = 168 Swedish speaking; Mage = 30.18, SD = 12.08, range = 18–72) provided 208 arguments against using hen. Most participants (88.7%, n = 149) provided one argument, whereas a minority (11.3%, n = 19) provided multiple critical arguments in their response. The analyses focus on the 208 arguments rather than participants.

Participants were recruited on the university campus of a large university in Sweden in 2014 (paper-and-pencil questionnaire) and through social media and other online platforms in 2015 (online survey). They completed the study without compensation. Participants indicated their gender in a free-text response to avoid normative gender categorizations (Ansara and Hegarty 2014; Lindqvist et al. 2019b), with 54.2% (n = 91) indicating woman; 40.5% (n = 68), man; and 5.4% (n = 9) not responding.

The study was completed in accordance with national guidelines on ethical research (Swedish Research Council 2017). In accordance with these guidelines, participants were informed about their voluntary and anonymous contribution and their possibility to quit the survey at any point without giving any reasons for quitting. They were also informed that results are presented on aggregated levels with no possibility to extract personal information. Next, participants gave their informed consent and were forwarded to the questionnaire. After answering the questionnaire, participants actively submitted their responses. A formal ethical approval was not mandatory for this type of research because it did not include any biodata nor did it intend to affect the participants physically or psychologically. It also did not involve any handling of sensitive data as described in the Swedish data protection law.

Coding

The presentation of the coding procedure and results follow guidelines outlined by Chatfield (2018) for publication of qualitative research in Sex Roles. To examine the content of the criticism of the gender-neutral pronoun hen in Sweden, we used a coding scheme based on Blaubergs (1980) and Parks and Roberton (1998). The initial coding scheme included 12 categories of arguments against gender-fair language. One category was excluded (Lack of Understanding) because it contained participants’ judgments of why others are critical of gender-fair language, which was not the focus of the present study. This initial coding scheme is reported in the online supplement.

The 208 negative arguments were coded through thematic analyses (Braun and Clarke 2006). The approach to the thematic analyses was both deductive and inductive. The deductive approach involved coding the arguments into the original categories, whereas the inductive approach involved the categorization of the arguments that did not fit into the original categories. First, two of the authors separately coded the arguments into the 11 original categories. The inter-coder reliability for coding into the original categories was indexed with Krippendorff’s alpha (Hayes and Krippendorff 2007) and was .89 for all categories taken together (range = .57–1). The average agreement between coders for coding arguments into the separate original categories ranged from 33.33% (this category consisted of only 3 arguments) to 100% (Magreement = 80.34%, SD = 23.4). The 15 arguments disagreed upon were categorized after joint discussion. This resulted in 45.7% (n = 95) of the arguments being placed in the original categories.

Second, the same two authors jointly subjected the arguments that were partly matched to original categories (35.1%, n = 73) to a second round of analysis, in accordance with how Parks and Roberton (1998) modified Blaubergs’ (1980) taxonomy. The goal was to decide whether these arguments could be included in the original categories if any of the category labels and category definitions were modified. The outcome from this procedure led to the inclusion of all arguments in modified original categories.

Third, the arguments that did not match any of the 11 original categories (19.2%, n = 40) were subjected to an inductive thematic analysis (Braun and Clarke 2006) by the same two authors. This led to the creation of two new categories and an updated coding scheme. The titles and definitions of the new categories were based on the explicit content of the comments grouped into this category. Subsequently, a third author coded all the arguments into the final coding scheme with 13 categories. In the third round of analyses, the Krippendorff’s alpha for the agreement between the third rater and the first two raters for the categorization of all comments was .98. The average agreement between coders for coding arguments into each of the original categories ranged from 66.67% to 100% (Magreement = 94.88%, SD = 11.08). The four arguments disagreed upon were categorized after joint discussion.

Finally, when all the comments were categorized, we used an inductive thematic analysis to organize the categories according to their latent content. To establish the themes, we went beyond the explicit content of the categories and searched for underlying ideas, assumptions, and ideologies that shaped the semantic content of the data (Braun and Clarke 2006). All authors were involved in this inductive process.

Results

The coding of the 208 negative arguments against hen into the categories showed that a majority of the arguments (n = 160, 76.9%) fit the categories in the original taxonomies from past reforms. Two original categories were modified: Sexism Is Acceptable was modified into Sexism and Cisgenderism Are Acceptable and Change Is too Difficult into Change Is Too Difficult or Unnecessary. About a fifth of the arguments (19.7%, n = 41) were coded into two new categories: Gender Identification Is Important (13.0%, n = 27) and Distractor in Communication (6.3%, n = 13). Two categories from the original taxonomy did not occur in this sample: Cross-Cultural and Historical Authenticity.

The inductive analyses of the latent content of the categories resulted in four dimensions of criticism of gender-fair language: (a) Defending the Linguistic Status Quo (39.4% of arguments, n = 82), (b) Sexism and Cisgenderism (27.4%, n = 57), (c) Diminishing the Issue and Its Proponents (26.9%, n = 56), and (d) Distractor in Communication (6.3%, n = 13). The dimensions and their underlying categories are presented in Table 1 with percentages and examples.

Table 1 Dimensions with subcategories, their definitions, examples, and frequencies
Full size table

Defending the Linguistic Status Quo

The first dimension captures a variety of categories that justify the current linguistic norm and the preference to keep the current linguistic system unchanged. Over a third of the arguments (39.4%, n = 82) fit into this dimension. This dimension included four categories: (a) Change Is Too Difficult or Unnecessary, (b) Appeal to Authority, (c) Word Etymology, and (d) Tradition. Change Is Too Difficult was modified to include the argument that change is not necessary. Comments about hen being unnecessary were very common, for example: “I don’t see the need for it” and “We already have a neutral word for gender non-specified situations: ‘person’.” Responses about the difficulty of use included: “I cannot get used to it” and “I have still not grown accustomed to the word and it is awkward to say.” Of the arguments, 33.7% (n = 70) fit into this category.

The Appeal to Authority (3.8%, n = 8) and the Word Etymology (1.0%, n = 2) categories rarely occurred and contained a variety of references to authorities. Some participants stated that they did not want to use hen because they found hen’s definition unclear or because they thought it lacked proper case forms. In the two comments in the Word Etymology category, it was argued that a generic use of hen is not proper because hen was introduced as a pronoun for people with trans or non-binary gender identities. The ambiguity in these arguments is that they are negative toward the generic use of hen, but not negative toward using hen for people who use it as their personal pronoun. Like the Change Is Too Difficult or Unnecessary category, the two arguments that fit into the Tradition category (1.0%, n = 2) showed a preference for the status quo. However, instead of referring to authorities, they used the tradition of using certain gendered words as justification for using them. For example, one argument stated that Santa Claus had always been a “he” and that should not change.

Sexism and Cisgenderism

The second dimension represents beliefs about gender and the implications of the language reform. The term “cisgenderism” was added to sexism and refers to the ideology that condemns people’s own designations of their genders and bodies (Ansara and Hegarty 2014). A fourth (27.4%, n = 57) of the arguments fit this dimension which comprises the modified category Sexist and Cisgenderist Language Is Acceptable (14.4%, n = 30) and the new category Gender Identification Is Important (13.0%, n = 27). In past taxonomies, criticism of gender-fair language included classical and hostile forms of sexism (e.g., “men are superior to women”). These forms of sexism were absent in the current sample, but we found similar hostility against people with non-binary identities. Trans identities were explicitly or implicitly neglected in the content of the comments (“I don’t see that a few hurt people should change language”). In other types of comments, the existence of people outside the binary system was neglected. According to these comments, pronouns for individuals with non-binary gender identities are unnecessary because there are no individuals who are not either a woman or a man: “I don’t see the need for a gender-neutral pronoun, because biologically you are either a man or a woman” or “Girls are girls, boys are boys.” Instead of an acceptance of masculine dominance being expressed in language, these comments express an acceptance of binary gendering in language. Therefore, we expanded the category Sexist Language Is Acceptable to Sexist and Cisgenderist Language Is Acceptable.

The new category Gender Identification Is Important emerged from comments focusing on the importance to use gender labels in communication about others and concerns that gender-neutral pronouns lead to depersonalization. This category also comprises dichotomous beliefs about gender. However, instead of explicitly denying the existence of individuals with non-binary identities, this category contains arguments about gendered information being important in communication (“I think one’s gender is part of who we are, and that is why I like saying he or she” and “one often wants to know someone’s gender”). In this way, the arguments implicitly neglect non-binary gender identities.

Diminishing the Issue and its Proponents

The third dimension includes disparaging reactions to both gender-fair language and the people advocating for its use. A fourth of the arguments (26.9%, n = 56) fit this dimension. This dimension comprises four original categories: (a) Hostility and Ridicule (12.5%, n = 26), (b) Freedom of Speech/Unjustified Coercion (9.6%, n = 20), (c) Sexist Language Is a Trivial Concern (3.8%, n = 8), and (d) Sexist Language Is Not Sexist (1.0%, n = 2). Common to these categories is the opinion that sexist language is a non-issue and that proponents of gender-fair language miss the larger picture of gender equality. The largest category within this dimension, Hostility and Ridicule, contains comments that undermine linguistic change by being hostile or diminishing: “It is totally ridiculous to use that word,” “It reminds me of a hen [referring to the English meaning of the word hen],” “It’s frivolous,” and “Seriously, I get irritated.” Clearly negative words were common in this category (e.g., ugly, nonsense, annoying, stupid, fussy).

Most of the ridiculing arguments in the first category target the word itself rather than the people using it. However, in other categories the arguments also target proponents of hen. For example, the Freedom of Speech/Unjustified Coercion category included criticism that focused on proponents of the linguistic change threatening or coercing others to change their linguistic habits. Examples of typical arguments are: “Sweden has ended up as a self-limiting totalitarian state where political correctness hides everything in society” and “It disturbs me that some abuse it, or even become aggressive if you decide to distinguish between the genders.” Some criticism was milder but suggests that gender equality had gone too far. An example of such a comment is: “I can feel [gender equality] has gone a bit too far if, for example, we will have to divide toilets for he, she and hen.”

The third category Sexist Language Is a Trivial Concern included arguments that express the idea that a word does not make a difference or that there are other more important steps to reach gender equality. The fourth category Sexist Language Is Not Sexist similarly included arguments that denied that sexist language negatively affects anyone. An example of such a comment is: “[Using gendered language] does not necessarily mean that certain characteristics or attributes are forced on a person.”

Distractor in Communication

The fourth dimension focused on participants’ concern that the use of hen may lead to a less effective communication or even reprimands because hen is seen as a political statement. This theme does not concern the content or the implications of hen, but rather the reactions from others to the new word. In total, 6.3% of the arguments (n = 13) fit into this dimension. For example, participants mentioned that hen could be distracting: “The reader will get stuck on the use of the word hen and it takes focus from the rest of what one is trying to say/write.” This category is similar to comments that diminish the issue and its proponents because it contains a “shoot the messenger” attitude. The difference is that the participants express concern that they themselves might become the target of hostility and ridicule when using hen because hen is perceived as a political statement: “I don’t want to use it at work during for example a presentation because there are many that have strong opinions about the word and I don’t want it to take the attention from the rest of the presentation.”

Discussion

The present study investigated whether criticism against gender-neutral pronouns in Sweden fit into taxonomies based on criticism of other gender-fair language reforms that aimed to make women more visible in English language contexts. The results showed that a majority of the arguments against the use of the Swedish gender-neutral pronoun hen could be categorized into existing taxonomies of criticism of gender-fair language (Blaubergs 1980; Parks and Roberton 1998). This similarity is noteworthy because the hen-reform differs from the reforms central in past studies in that hen removes gender information instead of adding paired gendered forms. This finding indicates that despite the differences in context and intentions of the gender-fair linguistic reforms, the contents of criticisms largely remain the same.

In addition to validating the work by Blaubergs (1980) and Parks and Roberton (1998), we established four dimensions of beliefs on gender and language that underlie criticism of gender-fair language reforms. The broadest dimension included arguments that defend the linguistic status quo. This theme reflects people’s tendency to be generally negative about anything that is new and to prefer to keep things the way they are (Bäck 2013; Bäck and Lindholm 2014; Jost et al. 2004). This phenomenon is called status quo bias (Jost et al. 2004; Samuelson and Zeckenhauser 1988). For instance, studies have shown that people preferred a candy that ostensibly has been sold a long time as compared to if it was presented as a relatively new candy (Eidelman et al. 2010). There is a similar resistance toward changes in language, where studies have found that new words and expressions are described as “ugly” or examples of “bad language” (Andersson 2001; Kotsinas 1996).

Some of the participants in the current study did not oppose the use of gender-neutral expressions per se (e.g., they were positive toward using terms like “the person”), but specifically opposed the use of hen. This indicates that these participants prefer existing words and may therefore overestimate whether words such as “the person” evoke gender-neutral or gender-balanced mental representations. In fact, such supposedly neutral words often carry a male bias (Bailey and LaFrance 2017). For example, in a recruitment situation, the supposedly neutral “the applicant” was found to have a male bias, whereas hen did not (Lindqvist et al. 2019). It is possible that people who express criticism with this underlying theme may be more prone to change their attitudes when hen loses its novelty or when made aware of the shortcomings of already existing neutral terms.

The dimension Diminishing the Issue and Its Proponents included the strongest negative attitudes and affect toward gender-fair language. Criticism in this dimension targets both the reform and the people advocating for the gender-fair language. The hostile comments prevalent in our sample reflect the heated public debate in the media and in comments on social media when hen was introduced in 2011 (Wojahn 2015). Criticism within this category may be so strong because it is rooted in ideology; what is important in terms of social justice, and what means should be used to attain social justice. For example, participants may perceive gender-fair language initiatives as an infringement on their freedom of speech. They may perceive initiatives like hen as prescriptive and a form of language policing. They may also relate hen to other social justice initiatives that they may oppose, such as initiatives to make public spaces such as toilets gender-neutral or non-binary inclusive or campaigns for non-binary people to self-determine their gender in official documentation.

Gender-neutral language may also lead to participants experiencing their own social identities to be threatened. For example, past research has shown that people with a strong female or male gender identity are more negative toward hen and use hen less often (Gustafsson Sendén et al. 2015). Challenging social gender identities by introducing a gender-neutral pronoun may be difficult for people and lead to harsh reactions (Morgenroth and Ryan 2018). Arguments that diminish the issue and its proponents may therefore be based on a person’s perceived threat of their own social identity.

The theme Sexism and Cisgenderism might also be rooted in ideology. This theme contained the arguments that gender-neutral pronouns are unnecessary because there are only two genders, that gendered language is important for one’s identity, and that knowing someone’s gender is necessary in communication. The content of this theme aligns with findings of prejudice against transgender individuals (Tee and Hegarty 2006). One predictor for prejudice against trans individuals, including individuals with non-binary gender identities, is the idea that sex and gender are determined by genes, hormones, and genitals (Tee and Hegarty 2006). Sex/gender is seen as unchangeable and binary, and transgender or intersex conditions are seen as abnormal. Classical forms of sexism that typically target women are not as relevant when proposing gender-neutral pronouns because they do not promote the visibility of women in language. However, it is possible that people who hold classical sexist beliefs also have essentialist beliefs on cisgenderism, similar in the way prejudice toward one minority group tends to spill over to other groups (e.g., generalized prejudice; Akrami et al. 2011). Our study suggests that criticism against gender-fair language may be rooted not just in the perception of gender roles in society, but also in cisgenderism. Whether cisgenderist motivations for being critical of gender-fair language are easy to change should be investigated in future studies.

The dimension Distractor in Communication concerns the motivation to not use gender-fair language because it is distracting. Arguments in this dimension addressed a motivation that has not been part of previous taxonomies of criticism—namely that participants are not necessarily critical of gender-fair language themselves, but instead are concerned about whether hen disrupts their communication with others and about whether other people may judge them when using hen. In a similar way, the American Psychological Association (APA) guidelines for authors support gender-fair language, but also recommend against using paired pronouns as they are “awkward and distracting” (pp. 74, APA 2009). Linguistic norms like guidelines and policies are important motivators for using gender-fair language (Koeser and Sczesny 2014). It is important that such guidelines are based on empirical evidence, and future research should investigate whether paired or gender-neutral pronoun forms truly are distracting in communication. It is also possible that the use of a word that is initially perceived as distracting or ideologically charged might normalize over time, as has been shown for paired masculine and feminine professional titles (Horvath et al. 2016).

Limitations and Future Research Directions

Many participants were students, who in general are more progressive and positive to hen as compared to the wider Swedish population (Bäck et al. 2018). The distribution of comments over the different categories and the themes may thus differ from the general population. A follow-up study could make use of a representative population to validate the results found in our study.

Future studies should investigate whether variation between individuals predict the type of critical arguments that are used. For example, high levels of sexism may predict hostile and sexist arguments, whereas a preference for the status quo may predict arguments concerning the preservation of the current language and already available words. Analyzing the dimensions of arguments over time may also reveal whether arguments with ideological motivations persevere longer than other arguments rooted in other motivations.

Practice Implications

Implementors of gender-fair language initiatives should be aware of the arguments that exist against gender-fair language to address resistance appropriately. Some arguments are based on ideological beliefs and values, which might be harder to change because they are associated with people’s social identities. Other arguments are based on the preference for status quo or a concern about making a political statement by using gender-fair language; these objections might change if the word becomes more common. Each of these underlying motivations may be addressed in different ways, and some may be more challenging to address than others. To facilitate the acceptance of new gender-fair language initiatives, it is important to provide knowledge about gender bias in language (Koeser and Sczesny 2014) and knowledge about variations in biological sex and gender identities (Hyde et al. 2018), as well as give people the chance to get used to the word so it loses its novelty. Guidelines for gender-fair languages for authorities and the media will help make the proposed word more familiar.

Conclusion

The introduction of the gender-neutral pronoun hen in the Swedish language has received a lot of criticism. In comparison to past gender-fair language reforms that made women more salient, this reform attempts to make people with non-binary identities visible in pronoun use and decreases dichotomous perceptions of gender. At large, critical arguments against using hen were similar to arguments against using paired forms. Four dimensions captured universal structures of hesitance to gender-fair language: (a) defending the linguistic status quo, (b) diminishing the issue and its proponents, (c) sexism and cisgenderism, and (d) gender-fair language being distracting in communication.