Gender inequality in academia from the perspective of the dialogical self: beyond ‘autonomous men’ and ‘relational women’

ABSTRACT The dichotomy of ‘autonomous men’ and ‘relational women’ is a long-lived social construction that is often taken for a fact. It is also suggested to relate to the reproduction of gender inequality in academia. Through the lens of dialogical self-theory, and based on our narrative data, we show that subtle gender inequality causes tension in the dialogical structure of the self. The tension is enforced by material and structural barriers and is rooted in the multiplicity of competing self-narratives. This tension is not an indication of lack of agency, and is linked to the interplay between self, organization and society. Those with a higher degree of multiplicity and democratic relations among their key I-positions are more prone to facing tension in career progression because of competing self-narratives. In this context, we introduce a distinction between ‘being autonomous’ and ‘having reduced multiplicity’. This distinction shifts the focus from women only to selfhood, and to men and the reduction of multiplicity of I-positions to address gender inequality.


Introduction
Studying gender inequality in academia from the perspective of the two main psychological aspects of selfhood (agency and communion) has been a mainstream perspective in diversity studies since the 1980s (Belenky et al. 1986). The key finding of this perspective is that the statistically observed agentic being of men and relational being of women alongside the pro-agency design of organizations is a fundamental contributor to gender self, organization and society, and is not an indication of lack of agency. If we do not acknowledge and problematize this tension we miss on explicating subtle gender inequality. On the contrary, if self is 'conceptualized as a vessel for storing all the particulars of the person' (Callero 2003) subtle gender inequality is disguised behind biased arguments based on agency as happening in postfeminist neoliberal argument. This paper contributes to gender studies by providing a thicker description of an intrinsic form of gender inequality in academia, which resides in the dialogues between self, organization and society. At a theoretical level this paper contributes to gender studies by promoting views on selfhood that consider relational ontology of the self and reject mainstream self-contained and individualistic views. This contribution is also offered by recent posthumanist studies on gender inequality (Taylor and Fairchild 2020). A message of the paper is that such views on selfhood are critical for revealing subtle gender inequalities, and the mainstream view on self is insufficient. Orienting the dialogical structure of the self at the core of the argument, this paper provides a steppingstone to move the gender discussion to a higher-order and gender-neutral discussion of selfhood, contributing to discussions of diversity and inclusiveness beyond gender.

Theoretical background
This qualitative study draws on extant narrative identity literature and dialogical self theory, which provide relevant context for its findings and a platform to expand on its discussion for theory building.

Narrative identity: agency and communion
According to the narrative identity literature, identity construction is shaped by the needs for agency and communion (McAdams and McLean 2013;Singer 2004). Consequently, these needs play crucial roles in self-narratives, helping us understand the perspective of a person. That said, narrative identity researchers are careful to orient this view of identity within the temporal and spatial context of poststructuralist Western societies and do not claim that it describes a universal phenomenon (McAdams and Olson 2010;Singer 2004). By viewing identity from this perspective, narrative identity researchers keep open the possibility of dynamic positioning of the identity and do not reduce it to certain psychodynamic forces (Singer 2004). McAdams and McLean (2013) describe agency as 'the degree to which protagonists are able to affect change in their own lives or influence others in their environment, often through demonstrations of selfmastery, empowerment, achievement, or status' and communion as 'the degree to which protagonists demonstrate or experience interpersonal connection through love, friendship, dialogue, or connection to a broad collective ' (McAdams and McLean 2013, 234).
Although there is no fixed law that agency and communion are gender specific, statistically, it is observed that it is more common for men to centre their identity around agency and women around communion (Abele and Spurk 2011;Gebauer et al. 2013). However, there are also studies indicating that this positioning can be dynamic over the life span (James et al. 1995). McMahon (2020), drawing on the works of Winnicott and Merleau-Ponty, shows conceptually how the cultural setup of our modern societies and nuclear families have created the dichotomy of agency vs. relational being between men and women, which is rooted in upbringings and societal norms. Hermans (2008) argues that the mainstream view of the self is in terms of 'self-contained individualism' and has misled us into considering the self a 'centralized unity with strict boundaries between self and environment that coincides with the skin'. As a contrast to this view, he proposes dialogical self-theory, which provides 'a conception of self and dialogue in which social power and dominance play a role in everyday life' (Hermans 2008, 191). In contrast to the individualistic self-view, dialogical self-theory posits a multiplicity of relatively autonomous I-positions within the structure of a self, each with a voice, that together 'result in a complex narratively structured self' (Hermans, Kempen, and Vanloon 1992). The Iposition should not be confused with the roles that a person has. These roles are social positions (e.g. I-as-teacher), while I-positions are the construal of personal positions within the self. The two may overlap in terms, but self-construal and its positioning with the self are the key point (e.g. I-as-ambitious, I-as-mother) (Hermans 2014) (Figure 1).

Dialogical self
Dialogical self-theory considers four key factors in the domain of self: 'the-other-in-theself, multiplicity-in-unity, dominance and social power, and innovation' (Hermans 2008). These factors allow us to attend not only to the coherence and alignment of narratives, Figure 1. Dialogical self as a complex of I-positions. Adapted from 'Self as a Society of I-Positions: A Dialogical Approach to Counseling', by Hermans (2014), Journal of Humanistic Counseling. which has been a common approach, but also to the multiplicity of and even competition between narratives (Bhatia and Ram 2001;Hermans 2001;Raggatt 2006). Lewis (2002) argues that although we can see multiplicity in earlier views of the self, for example, in those considering the ego, super ego, self-schema or narrative structures, they provide a static view of the self, while the self is a dynamically happening phenomenon, and considering the dialogue between voiced I-positions can reveal the dynamics.
Additionally, a critical feature of this theory for studying the self is that it takes into account both social power relations and personal agency simultaneously. Doing so puts this theory in line with the recent social constructionist views on the self, which addresses the epistemological gap between pragmatist and postmodernist views on the self. While the former highlights the role of agency and neglects the role of social power, the latter does the opposite (Callero 2003).

Research context and design
This research was designed to explore the subtle gender inequalities that operate in a less visible manner than overt discrimination (Monroe et al. 2008). Therefore, a qualitative study was chosen to investigate people's narratives in search of meanings in their selfconstrual and local knowledge about their selfhood, work and organization entwined with gender as a construct.
The selected case focused on the tenure process with an up-or-out design in an academic institute, where earlier statistics showed a lower completion rate for female tenure-track professors. This selection provided an example of an intense career progression process, making it suitable for research on the role of gender in career progression in academia. The study first aimed to determine the causes and viable explanations for the observed lower rate of progression of female tenure-track professors from the collected narratives. The findings were then analyzed at a higher order of theory building to arrive at a thicker description of the nature of gender inequality impacting women's career progression in academia.

Data collection
The qualitative study for this research project was conducted by interviewing 36 informants from both genders who had experienced the tenure-track process at academic institute X. The informants were selected from three categories, including those who had successfully completed the tenure-track process of institute X, those who had not, and those who were still going through the process at the time of the interview and were in the last two years of their up-or-out contract window. The interviews were semistructured with open questions to avoid leading the conversation to a certain conclusion and to minimize bias. The interviews were conducted in person or through video calls. Each interview on average took approximately 45 min. The interviews were designed to allow the interviewees to reflect on various domains from their professional and personal lives and share their perceptions, including: . Current situation, environment and others . Career, progression journey, organization . Self, professional and nonprofessional identities . Goals in life This design helped us to obtain temporally (past-present-future) and spatially (selfothers-environment) wide narratives for analysis.

Data analysis
Data analysis was conducted in a rigorous and iterative way. The interviews were analyzed through a three-level coding process following Gioia, Corley, and Hamilton (2013) methodology. At the first level of coding, each interview was coded separately by identifying statements from informants regarding their views on their experience of the progression track; their organization, work, roles and responsibilities; supporting and hindering factors; the role of gender; their personal and professional identities; and last but not least, their life prospects. The first-order coding provided us with provisional categories, which were then grouped across interviews during the second-order coding process to yield theoretical categories (Pratt, Rockmann, and Kaufmann 2006). In the second-order coding, we arrived at eleven key categories, which encompassed individuals' experiences of the progression journey. These elements reflect the nature of the job, environment, individuals' meaning making of the self, the profession, the organization, and luck. The third-level coding analysis was then conducted to bring these theoretical categories into more abstract theoretical dimensions (Pratt, Rockmann, and Kaufmann 2006). This part of the analysis aimed to find abstract dimensions, on which the theoretical categories from second-order analysis were arranged into a coherent picture. This resulted in five key dimensions, as shown in Figure 2, which could holistically illustrate the progression experience of individuals: effort and strategy, selfhood, socioeconomics, contextual influences and contingency. The first three elements capture three dimensions of an individual's self-experience: doing, being (or more precisely becoming) and having. These refer to what an individual actually does with respect to the demands of the progression, what she has become or is becoming through her personal and professional lived experiences, and what circumstances she has or had through the process (e.g. financial situation, mobility, pregnancy, parenthood). The fourth dimension is the organizational and social context surrounding the individual, which provides two-way external communication between the individual and the environment, resulting in different adaptations, influences, resistance, support, and obstacles over the tenure process. This also includes the material and structure barriers and gender biases. The last dimension is the inevitable share of contingency in a person's experiences. As we analyzed the interviews, the relevance of dialogical self theory and the narrative identity literature became more apparent, which created a perspective for elaboration of the findings that follow in the next section.

Interviewer:
Do you think gender plays a role? Int 5 (man, completed tenure track): … . I think the main thing could be that the males are generally a bit more competitive and that therefore this is a more competitive situation.
The nature of the work (meeting the criteria for the tenure-track process) was found to be a demanding process, requiring substantial effort from participants to meet the success criteria, including research publications, teaching, administrative tasks, grant submissions and student supervision. Therefore, putting in a significant amount of effort is a key requirement for success in this process. This includes but is not limited to time investment, time management, prioritization, stress coping mechanisms, ability to reach out for collaboration, having a realistic strategy for publication and teaching quality, and multitasking. We found that next to effort and having a suitable strategy, there were various degrees of contingencies or luck that played influential roles in some cases. This, for example, could take the form of encountering someone who later became an influential collaborator, having had a head start with publications from PhD or postdoc periods, having received quick responses from reviewers on submitted papers, or benefiting from the easier publication process in some fields relative to that in others. Considering that contingency is out of our control, the key drivers for success within participants' circle of influence are putting in effort and having a suitable strategy. We also found that the design of the progression system was construed as competitive and individualistic. Therefore, although effort and strategy are highlighted as the main elements in the career progression of our informants, looking through the lens of the 'big two' traits (agency and communion), we acknowledge the observation in earlier research that such a system favours autonomous being. Interestingly, we found that this lens appeared to be common among informants (as demonstrated in the sampled quotes below). However, our data reveal that this dichotomous characterization provides only a thin description and is an overgeneralized or even stereotyped view. Considering the multiplicity of self-narratives and the dialogical structure of selfhood, we arrive at a thicker description.
(2) Dialogical self: beyond the 'autonomous men' and 'relational women' dichotomy The collected data show the multiplicity of self-narratives arising from different temporal and spatial aspects of a person's lived experience. We also observe how these narratives are influenced contextually by the culture, environment and role of authorities.
Various key I-positions were highlighted throughout our interviews, such as I-asresearcher, I-as-teacher, I-as-mother, I-as-father, I-as-female, I-as-male, I-as-ambitious, and I-as-contributor-to-society.
Rather than assigning agency or communion at a universal level to a self, we can discuss agency and communion locally at the level of I-positions. The multiplicity of I-positions and the structure between them matter as well. A person who is construed as agentic is in fact being agentic in a certain I-position that has dominance over the other I-positions. For example, in the organization that we studied, a person with high autonomy in I-as-researcher, whose other I-positions are subordinate to this one, will find it easier to perform in the career progression setup of this organization. While a person who is construed as communal does not necessarily lack agency in the professional domain or specific I-positions, and this does not mean that the person carries only relational I-positions, such as I-as-contributor or I-as-helping-others, but is managing multiple I-positions simultaneously with more or less equal significance or some with even more significance than I-as-professional. In other words, they are relational within the structure of their selfhood. Hence, while the degree of agency should be looked at with respect to the task at hand, the degree of relational being should be looked at in relation to the structure of the dialogical self. This view on agency and communion takes us beyond gender divisions. (3) Competing narratives: the rise of tension in the dialogical self Different informants made different sense of the academic position, tenure track, and drive for success. To some, the autonomy of the tenure-track process was highlighted; to some, it appeared highly uncertain; and to others, it was a stressful and fearful period. To some people, the alternative if they did not achieve the tenure track was bright, offering a gateway to other opportunities in different industries or consultancies, while to others, the professional world outside academia, such as industry or consulting, seemed gloomy and was associated with hierarchy, having a boss or being in a prison with no autonomy. Similarly, we observed different perceptions of being an academic, such as being a learner, researcher, teacher, entrepreneur, contributor to society, or in a doubtful position lacking meaningfulness.
Competing narratives among I-positions cause tension in the self, hindering career progression. Tension can be present even when the individual has a high degree of agency in the critical I-positions driving career progression (e.g. I-as-researcher in this study). The tension becomes important when the demands of the job and the self-construed meaning of the job and the dominant I-positions are competing. For example, while Ias-researcher has considerable say in progression, I-as-contributor-to-society might be the dominant position in the person and have a conflicting narrative. Similarly, aligning narratives between I-positions can help the individual achieve set goals. In our analysis, we came across several competing narratives between the self-construal of 'my job' and 'I-as-academic', 'I-as-collaborator', and 'I-as-contributor'. Some example descriptions for 'my job' include the following terms: 'publication game', 'impactful', 'individualist', 'competitive', 'pointless', 'independence', 'entrepreneurial', and 'uncertain'. The two quotes below from two informants offer a good example of competing narratives.
Int 23 (woman): I realized I was writing papers that were not necessarily read so much … I still had two years and I only needed one paper to get the points, but I just didn't see the point anymore. I found a good opportunity elsewhere and then I left … Academia stopped having meaning for me. Int 6 (man): You publish all these articles, and nobody reads them anyway. … I know it doesn't really make much sense what I'm doing. Because really what I'm writing these articles for is a really select small group. But I like the game, I do. … I just like the game of publishing, of getting these publications out. And maybe, I can imagine that there's something that maybe appeals more to men than to women.
While publication is a large demand of the job, for the woman who left the organization, Ias-contributor had more dominance, and the tension between this and their self-construal of the meaning the job (pointless, meaningless) was high. In contrast, for the man who succeeded, despite having similar self-construal of the meaning of the job, his I-as-publisher had dominance. Beyond competing narratives between the dominant I-position and 'my job', mutually competing I-positions, such as I-as-mother and I-as-researcher, also lead to tension of resources (e.g. time, energy), which are to be shared between the two. The following quotes from two successful female and male tenure-track professors illustrate that for the woman, the tension between I-as-parent and I-as-academic was high, while for the man, I-as-academic had more dominance.

Int 4 (woman): That was the worst time in my life. I didn't enjoy my baby the way I should
have. The baby years, those three years, they just flew by, and I was so stressed out that I don't even think I enjoyed any of those moments that you are supposed to enjoy when you have a baby. Int 11 (man): I'm a person that has, I think, great self-efficacy. … I often have the feeling that whatever challenges you throw in my way, I can deal with [them]. I can make it successful. And that helps me a lot in seeing myself as a researcher and teacher because I know I can handle the pressures of the job … I worked countless weekends, and then, of course, my girlfriend was also maybe not so happy that I did that. I also have a daughter who would like to see me more often and spend more time with me.
(4) Gender in society and gender in self: external barriers, biases and internalized voices of others The rise of tension in dialogical self is not happening in isolation but in the interplay between self, organization and society. We found material and structure barriers impact women more than men and enforce this tension. This includes for example the tension between demanding period of pregnancy and limited time for tenure track. Another example is the tension between the need to relocate to another city (in case of not succeeding with the tenure track) and less mobility for women influenced by the culture in that there is less of an expectation for a man to compromise his career for his spouse than vice-versa.
Beside these external barriers gender biases in organization or society further enforce the rise of tension in dialogical self. During the interviews, the suggestion often came up that perfectionism is more common among women, keeping them from taking a realistic and good-enough teaching strategy as opposed to being perfect. However, from the dialogical self-perspective, for a person with a strong I-as-supporting-students, providing high-quality teaching is deemed a necessity because of its perceived meaningful impact on others, and compromising on this aspect can be construed as being inauthentic; hence, even if the person decides to compromise and save time for research still feels the tension.
Int 28 (woman): There's this tension that, on the one hand, it's great to spend time on teaching, but it also feels like it's eating up time, because of which you feel the pressure. … as a woman, I feel a bit more pressured to help out and to be nice. And it usually also feels good when you can do that, but it's difficult when you feel like you can't. So I've tried to not do it, but that also didn't really feel right.
Other examples of biases include biased teacher evaluations, or not being taken seriously as a professional during conferences, as stated in this quote: Int 27 (woman): there are these breaks at these conferences when people usually talk about their research and they do networking and so on. When I see two males talking, no matter how senior or how junior they are, they immediately start talking about research … But for example, if a senior professor approaches a female junior person, they start talking about weather or clothing, or so on. So this is something I observe and it really annoys me. Because I am also there, I have the same academic background and skills as the other junior male person has.
While gender biases exist outside and within society, we observe that they influence the narratives of the gender I-position that lives inside and plays a role in shaping the self-construal of the person. For example, fear of not being considered ambitious if seen as a mother. The following two quotes are examples demonstrating the internalized voices around the gender in self: Int 13 (woman): When people said, oh it must be so tough, I was always like, no, not at all. Because I didn't want to be seen as the mother. I want to be seen as a competent colleague and not as someone who now had a child so who was no longer ambitious. Int 9 (woman): You need to play the cards like men, and then they treat you like men, I guess. … I think sometimes life is easier for men. I don't really have this type of charisma. … when I have to teach in front of a crowd of 150 students, sometimes I think things would be easier if I was a man.
We also observed that I-as-mother is a bold, pronounced position, which is also influenced by the cultural context defining what acceptable motherhood is. As a result, working eighty percent was considered the norm, and deviating from it was construed by some as not being a good mother or created a sense of guilt. In general, such social pressure does not apply to fatherhood.
(5) 'My supervisor' as critical position: voice of authority A supportive supervisor seemed to be crucial for progression, and a supervisor holding gender biases was identified as having a clear and significant negative impact. Given the influence and authority of the supervisor in the organization, the narrative about the supervisor carries significant weight within the dialogical structure of the self. If this narrative is positive, it offers enormous support, and if it competes with other significant Ipositions, it can create great deal of tension.
Int 17: I never felt stereotyped or discriminated against during the tenure track. I also had a very supportive supervisor. … I never felt it was unfair, for example, against women. No, I never felt disadvantaged. Int 25: [my supervisor is] a good guy, but he really has this idea that when women have kids, there are these changes and they become lazy and demotivated. So, again, very old-school mindset, but in the end, it didn't help him because I left. And I am doing really well in business now.
(6) Gender inequality: tension in the dialogical structure of self-arising from dialogues between self, organization and society As a result of competing or conflicting narratives, tension arises within the self. This tension is linked to the dialogues between self, organization and society. As discussed above the external structure and material barriers and biases impact women more and enforce this tension. In addition to that, based on our analysis, our conclusion is that women are also more prone to heightened tension because of the more democratic structure of their selfhood, meaning that one I-position does not dominate the rest; for example, I-as-researcher, I-as-teacher, I-as-contributor, and I-as-mother can coexist with more or less equal significance. Additionally, and in contrast to I-as-male, I-as-femal faces gender biases, making it more prone to develop conflicting narratives with I-as-professional, which consequently adds more tension. On the other hand, men are more accustomed to holding I-as-professional in the dominant position. The idea that women are relational and men are autonomous is a thin and biased description for this common structuring of selfhood.
Int 13 (woman): Who am I today? … I am a mother and a partner and a friend and a daughter and a researcher and a teacher and a trainer and someone who is very eager to learn and to explore and really interested in doing that together with people.
The quote below from a man who succeeded with the career progression but faced tension between I-as-researcher and I-as-supporting-others is a good example of the tension linked to the dialogical structure of the self and not gender per se: Int 11 (man): Being able to say no more and to choose for yourself, to put yourself in first place and your research, instead of being collegial, providing more time to students who need more time. You become more self-centered in a way. For me, it was an uneasy feeling.
(7) Workload as an amplifier of gender inequality: the merger of professional and personal domains We found that the workload is such that professional tasks overflow into the personal domain, where for men, maintaining the dominance of I-as-professional is more common. In women, I-as-mother brings a significant competing narrative. As a result of this high workload, the interaction of self and society becomes critical. Hence, workload is a crucial element in influencing gender inequalities, albeit indirectly. Becoming or being a mother with its cultural context for some individuals has a greater impact on their work-life balance than can be easily managed by the amount of maternity leave provided. Pregnancy, delivery of a baby, and caregiving in the early years can pose physical, psychological or financial demands that compete for shared resources (the person's time and energy). The demands of the tenure track are such that people often need to invest in work from their personal time, and depending on their means and situations for managing the demands of pregnancy and motherhood, divesting from their personal time to invest in their professional role can be less affordable for some than it is for others. Although many managed, our observation is that they bore a higher level of tension. Even in some cases where the father provided more support than the mother in parenting, motherhood remained a pronounced and demanding role.

Discussion
The purpose of this paper was to investigate whether gender inequality still plays an adverse role in the career progression of women in academia and, if so, what the nature of this gender inequality is. Based on the findings of this study, we arrive at an explication with three postulates on gender inequality. First, subtle gender inequality arises from the dialogues between self, organization and society. Second, looking through the lens of dialogical self-theory and attending to the multiplicity of I-positions provide a thicker description of this subtle gender inequality, going beyond the limiting dichotomy of 'autonomous men' and 'relational women'. Third, the nature of this subtle gender inequality is tension in the dialogical structure of the self, which is not rooted in individualized psychological characteristics, but in interplay between self, organization and society.
The first postulate echoes the conclusion of some recent studies on gender inequality pointing out the multilayered nature of this phenomenon taking place in the interplay between individual, organization and society (Erdogan 2021;McCarthy and Moon 2018). Along with gender in organizations, gender in self and gender in society must be considered simultaneously; for instance, Greguletz, Diehl, and Kreutzer (2019) study the root causes of less effective network building by women and reveal the importance of the aspects of gender in self and society (self-hesitation, moral social interaction). Savigny (2014) highlights the normalized sexism in culture and society (gender in society) as key elements adversely impacting women in academia (gender in organizations). As opposed to overt gender inequality, this form of gender inequality remains subtle because no single, definite factor can sufficiently explain the phenomenon, which instead lies in the dialogue between self, organization and society. Without attending to this dialogue, we fall short in exploring a thicker explication of this phenomenon and the phenomenon is construed as 'it is what it is'.
The second postulate points out that the abovementioned interplay between self, organization and society becomes observable in the dialogical structure of the self. Attending to the multiplicity of self-narratives and dialogical structure of the selfreveals how resources (time and energy) are allocated among significant I-positions and how much tension there is between competing or conflicting narratives. Although career progression is a professional matter, attending only to the narratives of this domain limits our understanding of the full picture, especially in relation to gender inequality. Considering one I-position (e.g. I-as-publisher) in isolation from others could give the false impression that person A is more agentic than person B. However, considering their other I-positions and how they are structured in relation to each other could reveal that person B is not less agentic in that particular I-position but that meeting the demands of the organization creates more tension in person B.
The third postulate shows the nature of subtle gender inequality by pointing to the creation of tension in selfhood. I-positions can be in alignment, conflict, or competition. The internal I-positions are not only in dialogue with themselves, but also with the external world, the organization and society. Structure and material barriers and gender biases are key external sources that impact the women more and enforce this tension. Workload also plays a critical role in this tension. Once the workload is such that the professional domain trespasses into the personal domain, conflicts between the self in organization and the self in society or professional and nonprofessional I-positions are heightened, causing greater tension for women. Additionally, while the gender role and the biases associated with are outside, the gender I-position is inside, influenced by them, and plays a role in shaping the self-construal of the person. This is a less studied aspect of gender inequality that influences the internal perception of one's own ability and career choices (Orupabo 2018). Because of these internalized voices, I-as-female and Ias-mother have more sensitivity around them than I-as-male or I-as-father. Hence, women are more prone to facing tension.
Based on this explication of gender inequality, we argue that individuals who reduce or suppress the multiplicity of their I-positions such that the key I-position for career progression assumes the dominant role have a higher chance of success, while those with a higher degree of multiplicity and more democratic relations between their I-positions are more prone to facing tension. Although these go beyond gender and are a matter of the dialogical structure of selfhood, it appears that the former (a reduced multiplicity of I-positions or a hierarchical structure of the self) is more common in men, while the latter (a higher multiplicity of I-positions and a democratic structure of the self) is more common in women. Following this line of thought, we argue that what is perceived as autonomy in men could actually be the reduction or suppression of multiplicity in a hierarchical structure of selfhood. Such a structuring of selfhood does not suit everyone and runs counter to the concept of the dialogical self, which considers a democratic dialogical self crucial for imagination and relational being (Hermans, Kempen, and Vanloon 1992).

Future research
Based on the findings of this study, three proposals for future research are suggested.
The first proposal is to go one step further and question the dichotomy of autonomy and relational being. We propose thatinstead of considering these two terms as a dichotomywe work with the single term degree of relational being, and save the term degree of autonomy for the interaction of the self and system or task. This proposal also resonates with the outcome of the quantitative study by Gebauer et al. (2013) on selfesteem, showing its relation to the harmony between activity and personality rather than merely personality.
The second proposal is to incorporate the intersections of race and class with gender in studies on gender inequality, as proposed by recent studies, but still from the perspective of the dialogical self. Bringing I-positions associated with class and race might further reveal the tensions in selfhood. According to some scholars, postfeminist academia has tried to address gender inequality through elitism by giving privilege to white Western women from the middle or upper classes, while women from marginalized races and classes not only face more inequality but are also less cared for, since by presenting this elitist equality as true equality, the dominant neoliberal paradigm in academia is shifting the argument from gender inequality to a lack of personal agency (Gill, Kelan, and Scharff 2017;Liu 2019;Rucker, Galinsky, and Magee 2018).
The third proposal is to shift the focus from women to men and study the construction of selfhood in men in response to power relations and social expectations in multilayered investigations on gender inequality. Maleness has remained unproblematized in gender studies, while interrogations of femaleness have been a common trend in this literature (Runte and Mills 2006). According to rising critics of postfeminist neoliberalism, academia has closed its eyes to social barriers and individual differences while demanding overwork and performance in an individualistic and competitive way (Gill, Kelan, and Scharff 2017;Liu 2019;Taylor and Gannon 2018). This study gives us some hints that reducing multiplicities of I-positions or imposing a hierarchical structuring of selfhood, whereby all I-positions are subordinated to I-as-professional could be a common way of responding to organizational demands among men.

Conclusion
Narratives of men being autonomous and women being relational are long-lived social construction shaping our construal of gender in an individualistic view of selfhood. Such narratives are enablers of gender inequality, and attending to dialogical structure of the self reveals their biases. In this study, we showed how subtle gender inequality leads to rise of tension within dialogical structure of the self. Material and structural barriers and gender biases also further enforce this tension. This tension is no indication of lack of agency.