Introduction
Over the past decades, academics across Canada have devoted a lot of attention to diversity and representation. This is particularly true for the Canadian Political Science Association (CPSA) which is well aware of the bilingual and multicultural nature of the country. Indeed, there are good arguments to favour a fair representation of minorities and under-represented groups among the discipline (e.g., Abu-Laban, Reference Abu-Laban2017; Vickers, Reference Vickers2015). The CPSA fosters several initiatives in that regard, including useful resources for academics around the country who lecture Indigenous politics, a women's caucus to promote equal opportunities and provide gender-based mentorship, and the flagship journal of the association (Canadian Journal of Political Science/Revue canadienne de science politique) recently agreeing to publish the annual presidential address in both English and French, which contrasts with the previous 104 presidential addresses published in English only (compared to two in French, in 1952 and 1975).Footnote 1
Despite many efforts, academics belonging to specific groups—including people of colour, non-native English speakers, women and Indigenous peoples—are under-represented at all stages of their careers, according to a variety of indicators. For example, they face more barriers to get postgraduate degrees (David, Reference David2015), and their work is less frequently cited (Jones and Floyd, Reference Jones and Floyd2024) and published in top journals (Teele and Thelen, Reference Teele and Thelen2017).
In this research note, we focus on the representation of French-speaking scholars within Canadian political science. While all under-represented groups should be thoroughly analyzed, there are good reasons to study linguistic divides in the Canadian context. First, bilingualism is a key feature of Canadian politics and Canadian identity (e.g., Chouinard, Reference Chouinard2021; Schwartz, Reference Schwartz1967). Second, the linguistic divide between English and French speakers has long been known to be salient within Canadian political science, leading to debates surrounding the language used during events like the annual presidential address (Abu-Laban, Reference Abu-Laban2017; Noël, Reference Noël2014) and whether members of juries evaluating book awards should be required to have the necessary linguistic skills to understand the books they review (Blattberg, Reference Blattberg2017). Third, according to studies on this linguistic divide, French-speaking scholars face under-representation, and their work is not as valued as research from English-speaking scholars. For example, they are “systematically absent or under-represented in the assigned readings of Canadian politics courses” (Daoust et al., Reference Daoust, Gagnon and Galipeau2022: 898; see also Rocher, Reference Rocher2007). All in all, “the work of French-speaking Canadian political scientists is poorly considered in the research and teaching of English-speaking colleagues” (Lecours, 2023: 781).
More specifically, our note focuses on the linguistic composition of panels and papers at the most important in-person event for Canadian political science, that is, the annual CPSA meeting. This meeting is a structuring event in many ways. In addition to providing networking opportunities, some of which are sponsored with the very objective of helping less privileged participants, scholars are exposed to the most up-to-date work in the field. Yet, the event's settings can lead to the (re)production of inequalities, especially if activities during the annual meeting are clustered around homogenous groups.
To examine linguistic diversity within Canadian political science, we generated a dataset based on the official program of the 2023 annual meeting. To the best of our knowledge, there is no systematic, quantitative analysis of linguistic diversity and representation during such a major event in Canadian political science. This stands in sharp contrast with the United States, where meetings of the national political science association have been systematically analyzed (Goodman and Pepinsky, Reference Goodman and Pepinsky2019; Gruberg, Reference Gruberg2006, Reference Gruberg2009; Piscopo et al., Reference Piscopo, Xydias, Atchison and Och2023). For our purposes, the 2023 CPSA annual meeting is particularly interesting given that it was the first in-person one since 2019 (due to the COVID-19 pandemic), with people eager to participate and get together to initiate new projects. Overall, the fact that this is the most important annual conference in Canadian political science, with potentially major implications on the (re)production of inequalities, makes it a relevant case study to assess the inclusiveness of the discipline regarding linguistic diversity.
Our main findings are twofold. First, we show an under-representation of events held in French and of French-language papers, which is in stark contrast with the baseline proportion of French-speaking scholars attending the event. While not necessarily surprising, the magnitude of the gap is astonishing. For example, about 94% of French-speaking scholars presented their work in English during the meeting, compared to 0.3% of English-speaking scholars presenting in French. Second, we measure linguistic concentration in panels and papers using the Herfindahl-Hirschman Index (HHI), which reveals that francophone-dominated panels and co-authored papers with francophone first authors are significantly more linguistically diverse than anglophone-dominated panels and papers. In other words, there is very little linguistic diversity at CPSA's annual meeting, and this lack of diversity does not result from francophones’ self-isolation. We conclude this research note by discussing the implications of our findings which, among other things, highlight linguistic divides and biases in Canadian political science and help make sense of the lack of representation of French-language work in courses on Canadian politics within universities, and in Canadian academia more broadly.
Linguistic Cleavages in Canadian Political Science
Bilingualism is an important feature of the Canadian polity. Beyond its strongly symbolic value, bilingualism, and more specifically the representation of French-speaking Canadians, has historically been at the very heart of Canadian politics (Gagnon, Reference Gagnon2014; Vipond, Reference Vipond1991). Put simply by Schwartz (Reference Schwartz1967: 83), “the most critical question that has arisen out of the complexity of social, political, and economic interests in Canada has concerned the relations between the French and the English.” While it used to be more numerous, the French linguistic minority still constitutes more than one-fifth of the overall population, and language continues to play a key role in Canadian politics (e.g., Bélanger et al., Reference Bélanger, Daoust, Mahéo and Nadeau2022; Brie and Ouellet, Reference Brie and Ouellet2020; Daoust and Jabbour, Reference Daoust and Jabbour2020; Brie and Mathieu, Reference Brie and Mathieu2021). Moreover, debates about linguistic representation punctually resurface in the political sphere, for instance, with the Superior Court of Quebec recently agreeing to study a complaint about Canada's current Governor General Mary Simon not speaking French, which would, according to the plaintiffs, violate the linguistic rights of francophone Canadians.
As a field of study, Canadian political science strives to be inclusive and, in many ways, values linguistic diversity. Among other things, the Canadian Political Science Association, just like the Canadian Journal of Political Science/Revue canadienne de science politique, is officially bilingual (Cardinal and Bernier, Reference Cardinal and Bernier2017; Marland, Reference Marland2017). However, French-speaking scholars face significant obstacles over the course of their careers. As detailed by Jean Laponce (Reference Laponce2006), English is the lingua franca in personal contacts between English- and French-speaking individuals, which is neatly illustrated by the notion of “asymmetrical bilingualism” (Simeon and Cameron, Reference Simeon, Cameron, Cameron and Simeon2009). In other words, francophones face a “forced choice” to communicate in English in professional settings (May, Reference May, Kymlicka and Patten2003).
This notion of “forced choice” has been strengthened by the strong structural incentives that junior scholars face if they want to maximize their chances of being hired by a Canadian university. That is, one must recognize that scholars have an agency in selecting the language they communicate their research in or the language that will be used during an event that they organize, but also that this agency is shaped by various structural forces. For example, one's decision to submit a panel in French comes with a potential limitation regarding the number of attendees; writing a full paper in French instead of English means that it will be submitted to a journal that is either francophone or bilingual, which are not among the most highly ranked in the discipline; it also entails receiving fewer citations for their work (Godbout, Reference Godbout2017), which is a key metric for both general prestige as well as promotions; French-speaking scholars might also want to present in English because they conduct their research in English, which makes it easier for them to present in that language; etc. In other words, scholars have an agency within a broader context in which structural incentives (for individuals) favour English (Di Bitetti and Ferreras, Reference Di Bitetti and Ferreras2017).
In this context, French-speaking scholars overwhelmingly present their work and publish in English (Blais, Reference Blais2018; Stockemer and Wigginton, Reference Stockemer and Wigginton2019). This phenomenon is most salient for junior scholars (Rocher and Stockemer, Reference Rocher and Stockemer2017). That said, even if the work of French-speaking scholars is widely accessible in English, francophones are still under-represented in Canadian political science. For example, Rocher (Reference Rocher2007) found that the knowledge production of francophone scholars during the 1995-2005 period was very rarely incorporated in English-language articles or books on Canadian politics, raising doubts about the inclusivity of political science. More recently, Daoust et al. (Reference Daoust, Gagnon and Galipeau2022) showed that despite mostly publishing in English, French-speaking scholars were strikingly under-represented in Canadian politics’ syllabi at both undergraduate and graduate university levels, to the point where the typical syllabus includes no French-speaking authors. Both works conclude that francophone scholars are under-represented because of systemic discrimination, which is likely to result in significant professional prejudice.
Based on these findings, we suggest that the under-representation and, in many instances, the exclusion of voices and concerns of one of the two core linguistic groups in Canada deserves to be more thoroughly analyzed. Indeed, in addition to normative arguments related to diversity, francophone scholars are likely supplementing the field with unique contributions due to their different cultural background and socialization. Evidently, these arguments also apply to other under-represented groups (e.g., Ladner, Reference Ladner2017; Nath et al., Reference Nath, Tungohan and Gaucher2018), and while we focus on linguistic under-representation, we hope that more work will be done to study other aspects of representation in Canadian political science.
To better understand the French language's representation in Canadian academia, we focus on the annual meeting of the Canadian Political Science Association, which has the potential to highlight how certain practices within Canadian political science can bolster or hinder linguistic inequalities. We detail our methodological approach in the next section.
Data and Indicators
We generated a new dataset based on the official program of the 2023 Annual Conference, held at York University from May 30 to June 1, 2023. More precisely, we collected information about the number of panels, the subfield (i.e., section) of each panelFootnote 2, the number of participants in every panel, the name of the participants and their respective roles (presenter, chair, discussant), the number of authors per paper, and most importantly, the language of the panelFootnote 3 and of the presented papers. An aggregate summary of our findings is presented in Table A1 (in the Appendix), which should be interpreted in light of the baseline mother-tongue distribution of participants presented in Figure 1.
Linguistic Distribution Panels, Papers, Authors and Chairs/Discussants
Note: Figure 1 presents the percentage of English- and French-speaking panels (by panel title language), papers (by paper language), authors (by mother tongue) and chairs/discussants, including co-chairs and co-discussants (by mother tongue). Participants with a mother tongue other than English or French are excluded in Figure 1. Each occurrence of the same individual is coded as a separate observation. N=193 panels, 673 papers, 924 authors, 380 chairs and discussants.
Each participant's mother tongue is coded as English, French or other. To obtain this information, we first used three types of information: (1) linguistic abilities as listed on participants’ resume, (2) language of undergraduate studies and (3) first and last name. Then, for cases where we were not entirely certain about one's mother tongue after aggregating these three types of information (less than 10 percent of authors), we validated our information with colleagues in our network. Using this data, we measured the linguistic distribution of panels, papers, authors and chairs/discussants, as well as the percentage of participants involved in research activities (papers and panels) in a language different from their mother tongue.
To assess linguistic cleavages in papers and panels, we computed the Herfindahl-Hirschman Index (HHI), a common measure of ethno-linguistic concentration that ranges from 0 to 1, where 0 means total fragmentation (all participants have a different mother tongue) and 1 means total concentration (all participants have the same mother tongue). The Herfindahl-Hirschman Index is now widely used to assess linguistic diversity quantitively within countries, especially for research on ethnic conflicts (see Bleaney and Dimico, Reference Bleaney and Dimico2017 and Laitin, Reference Laitin2000). It is also frequently employed to measure ethnic, language or gender diversity in the corporate world (Michie and Ougthon, Reference Michie and Oughton2022; Upadhyay and Zeng, Reference Upadhyay and Zeng2014) or within educational institutions (Widiputera et al., Reference Widiputera, De Witte, Groot and van den Brink2017). It is calculated using the following formula:
where S is the share of participants from each linguistic background (English, French or Other) within a paper or panel, and N is the total number of linguistic groups (i.e., 3). When normalized to range from 0 to 1, a HHI above 0.25 indicates a high level of concentration. In Canada, the baseline level of linguistic concentration (when considering the mother tongue distribution between anglophones, francophones and allophones) is 0.42.
Results
We begin by examining the linguistic distribution of panels (by panel language), papers (by paper language), authors (by mother tongue), as well as chairs and discussants (by mother tongue) during the 2023 CPSA conference. Figure 1 displays these findings and shows a clear under-representation of French-language panels and papers (respectively 1.6% and 4.2%) when compared to the baseline proportion of French-speaking authors and chairs/discussants (respectively 18.3% and 19.2%). Indeed, in the context of this research note, our baseline is the proportion of participants from each linguistic group at the 2023 conference, as we do not make a broader argument on differential participation rates to the conferenceFootnote 4. The density distribution of the percentage of French- and English-speaking panels and authors is presented in Figure A1 of the online Appendix, further showcasing the existence of unequal representation between both linguistic communities. The aggregate average of francophone participants across panels is 16.9%, with 45.1% of all panels having no francophone participants and only 2.1% of all panels having no anglophone participants. The aggregated average percentage of francophone authors across all papers is 15.9%, with 81.3% of all papers presented at the conference having no francophone authors and 23.2% having no anglophone authors.
One of the consequences of the under-representation of French-language panels and papers is the unequal distribution of non-native participations within the activities of the conference. Table 1 presents the percentage of English and French speakers involved in panels and papers in each of Canada's official languages. The overwhelming majority of francophones take part in English-language panels (93.5%) and papers (83.5%), contrary to less than 1 percent of anglophones doing so in French (0.3% for panels and 0.7% for papers). This echoes Simeon and Cameron's (Reference Simeon, Cameron, Cameron and Simeon2009) principle of “asymmetrical bilingualism,” according to which the burden of communication is often carried by francophones in French-English exchanges within Canada. Indeed, low levels of functional bilingualism outside the province of Quebec (9.5% in the Rest of Canada, versus 46.4% in Quebec, according to Statistics Canada, 2023) contribute to an unequal distribution of non-native work across Canada's two solitudes. Moreover, Figure A2 (of the online Appendix) presents the percentage of French-language authors, as well as chairs and discussants, across all subfields and other event categories. Among other specificities, it is startling to note that none of the special event roundtablesFootnote 5 featured a francophone panellist or chair/discussant in 2023. Out of all 14 subfields, nine show a percentage of francophone authors lower than the overall percentage of French-speaking presenters in the conference (i.e., 18.3%), while five have an over-representation of francophones compared to that baseline (Canadian Politics, Comparative Politics, Local and Urban Politics, Political Behaviour/Sociology and Provincial and Territorial Politics).
Language of Participants, Panels and Papers
Note: Table 1 presents the distribution of participants involved in panels and papers in French and in English by mother tongue. There are 190 English-language panels, 3 French-language panels, 515 English-language papers and 23 French-language papers. This data excludes chairs and discussants. Each occurrence of the same individual is coded as a separate observation. Data: CPSA 2023 Annual Conference Program (n=841 panel participants, 720 authors).
We now turn to the linguistic concentration of panels and co-authored papers presented during the conference. As previously mentioned, we use the Herfindahl-Hirschman Index (HHI)—a widely-used measure of ethno-linguistic fractionalization—to estimate the level of linguistic concentration within panels and papers (including languages other than French or English) with a score ranging from 0 (complete linguistic fractionalization) to 1 (complete linguistic homogeneity). To compare the linguistic diversity of panels and papers that are French- and English-dominated, we generated a binary variable assessing linguistic dominance. Linguistic dominance was defined by the majority mother tongue language (>50%) of all participants for panels and by the mother tongue of the first author for papers (for co-authored papers only). Our goal is to assess self-isolation as a potential cause for francophones’ under-representation within the conference. In other words, in this section, we test whether French-speakers have a stronger propensity to work with members of their own linguistic group than their anglophone counterparts do—and demonstrate that this is not the case.
Figure 2 (a) shows the distribution of the HHI score across all panels by linguistic dominance, revealing a significantly (p=0.002) higher HHI score for English-dominated panels (0.77) than for French-dominated panels (0.62). These results are confirmed by the linear regression model presented in Table 2, which controls for the total number of presentations when assessing the relationship between panels being French-dominated and their level of linguistic fractionalization (β=−0.15).
Linguistic Concentration in Panels and Papers by Linguistic Dominance
Note: This Figure presents the level of linguistic dispersion as measured by the Herfindahl-Hirschman Index (0=not concentrated, 1=fully concentrated) by panel and by papers, for each dominant language. Dominant language is measured as the majority language (>50%) in each panel based on the mother tongue of the authors, the chair, the co-chair, the discussant and the co-discussant (a) as well as by the mother tongue of the first author listed on each co-authored paper (b). Panels with a dominant allophone language (n=5) or with no majority language (n=20) are excluded from the analysis as well as co-authored papers whose first author is neither a French- nor an English-native speaker (n=11). Student's t-tests show that French-language dominant panels (p=0.002) and papers (p=0.010) are on average more linguistically diverse than English-language ones. The mean HHI for French-language panels and papers is respectively 0.62 and 0.81 and respectively 0.77 and 0.93 for English-language panels and papers. The overall HHI average for panels and papers is respectively 0.76 and 0.88, with a median of 0.76 and 1.00 and a standard deviation of 0.203 and 0.202. Data: CPSA 2023 Annual Conference Program (n=165 panels, 161 co-authored papers).
Panel Characteristics as Determinants of Herfindahl-Hirschman Index in Panels
Note: This Table presents results from a linear regression model testing the correlation between observing mostly francophone participants on a panel (dummy variable) and the total number of presented papers with the dependent variable (i.e. the Herfindahl-Hirschman Index of linguistic concentration within that panel). Panels with a dominant allophone language (n=5) or with no majority language (n=20) are excluded from the analysis. * p<0.1, ** p<0.05, ***p<0.01. Data: CPSA 2023 Annual Conference Program (n=165 panels).
Figure 2 (b) presents the HHI score for co-authored papers by linguistic dominance, revealing a similar trend, which is also significant (p=0.010). Indeed, the average HHI for co-authors with an English-speaking first author is 0.93, compared with 0.81 for those with a French-speaking first author. Especially telling is the median HHI value for English-dominated papers, which is equal to 1, indicating a perfectly homogeneous English-speaking concentration. To better grasp the effect of having a francophone first author on a paper's overall linguistic concentration, we conducted an OLS regression which controls for the total number of authors in papers. Results are presented in Table 3 (model 1) and show a statistically significant negative effect of having a francophone first author on a paper's overall linguistic concentration (β=−0.10). In other words, a paper is on average less homogenous (i.e., more diverse) when it has a francophone first author. Table 3 (model 2) presents an alternative test evaluating the province of academic affiliation of the first author as a determinant of linguistic concentration, showing that Quebec-based academics (β=−0.13, p<.05) and academics based outside of Canada (β=−0.12, p<.1) have on average lower HHI scores for authorship composition, indicating higher levels of linguistic diversity in their collaborative work, than those based in predominantly English-speaking Canadian provinces.
First Author Characteristics and Herfindahl-Hirschman Index in Co-Authored Papers
Note: This Table presents results from a linear regression model testing the correlation between having a paper's first author being francophone (dummy variable), the total number of co-authors on a paper and the province of the first author's university affiliation with the dependent variable (i.e. the Herfindahl-Hirschman Index of linguistic concentration on that paper). Alberta is the category of reference for the first author's provincial affiliation. * p<0.1, ** p<0.05, ***p<0.01. Data: CPSA 2023 Annual Conference Program (n=172 co-authored papers).
Conclusion
Despite Canadian political science striving to become more inclusive, several minority groups still suffer from important barriers. In this research note, we focus on linguistic divides within the discipline by leveraging the most important annual meeting for political scientists in the country: the CPSA conference. We show that there is a significant and substantially very large under-representation of French-speaking events and papers at the annual CPSA meeting, which suggests that this event remains largely oblivious to Canada's bilingual nature.
This directly results in French-speaking academics bearing the weight of collaborative work to allow for interaction with their English-speaking peers, as revealed in the near-absence of English speakers taking part in French-language panels and papers. We also computed Herfindahl-Hirschman indexes measuring linguistic diversity of panels and papers during the event. The indexes demonstrate that panels dominated by francophone participants and co-authored papers with francophone first authors are significantly more linguistically diverse than anglophone panels and papers. This raises interesting questions: who collaborates with whom (and why) within Canadian political science? And is the lower level of functional bilingualism in the Canadian English-speaking population (Statistics Canada, 2023), and potentially of anglophone participants at CPSA, the main driver of this gap? We hope that scholars will address these questions in future research.
Speaking about behaviouralism in Canada, Reg Whitaker wrote that a “historian innocently leafing through the Canadian Journal of Political Science might well wonder at times if political scientists have given up the use of both official languages and have taken instead to communicating in numbers alone” (Reference Whitaker1979: 1). This quote may come across as grim, but the dilemma in the 2020s context is even worse, as it is quite clear that political scientists have given up on bilingualism in practice. Canadian political science has been divided on many issues, but its strong commitment to diversity, as well as an intersectional approach to inequalities, seem to be quite consensual. Yet, diversity in the discipline encompasses linguistic diversity, and hence, the findings from this research note showing a substantial under-representation of the French language are important. The fact that most of the socialization processes of Canadian political scientists take place in the only “real” language of the CPSA (Noël, Reference Noël2014: 650), that is, English, is problematic.
What can be done to improve the linguistic imbalance during the annual meeting? One obvious action is to keep documenting the linguistic makeup of our events. This research note helps in that regard, but administrative officers, who are probably aware of the linguistic imbalance we describe, should also keep a yearly record of the proportion of events presented in French during the CPSA conference. However, this research note is meant to generate discussion within the discipline and ultimately raises more questions than it provides solutions. For instance, what is the role of section chairs in enhancing linguistic diversity? Should CPSA itself send a more explicit signal in the upcoming editions of the annual meeting that events in French are particularly welcome, in the hopes of bridging this linguistic gap? Moreover, the precedent set by André Lecours in 2023, who delivered his presidential address in English and French, is a significant symbolic gesture and raises the following question: should the presidential address always be delivered in the association's both official languages? Again, we do not provide an answer to these questions, but hope that relevant institutions and individual scholars will engage with these issues.
Finally, we are hopeful that participants themselves will be more conscious of the linguistic under-representation we documented in this research note. While structural incentives will likely remain strong for French-speaking academics to pursue much of their participation in English, they might be more inclined to submit papers or panels in French since it is now clear, as demonstrated by facts and figures, that French is severely under-represented during the event. Overall, as a discipline, English speakers and French speakers alike need to be more aware of patterns of linguistic under-representation and strive to improve the presence of French within Canadian academia.
Supplementary Material
The supplementary material for this article can be found at https://doi.org/10.1017/S0008423924000222.
Competing interests
The authors declare none.
Acknowledgements
The authors would like to thank Alain G. Gagnon for sharing his insights on a previous version of this manuscript. We are also thankful to research assistants Sofia Ouslis and Adrian Facciponte at Western University for their work on the 2023 CPSA dataset.
Open access
