How to Problematise Categories: Building the Methodological Toolbox for Linguistic Reflexivity

Abstract

Following qualitative researchers’ growing interest in reflexivity, a body of scholarship has emerged that aims to turn informal practices for reflexivity into methods that can be learnt and taught alongside other research practices. This literature, however, has focused on helping researchers become more reflexive toward their situatedness and positionality, rather than toward their use of language and its effects on knowledge production – a process I refer to as ‘linguistic reflexivity’. This article addresses this gap by formalising a method for ‘problematising categories’, an informal approach familiar to qualitative researchers as a promising solution to the analytical and ethical blinders that result from scholars’ unconscious use of language. I proceed in three steps. First, I review the literature to show the analytical, empirical and ethical rationales behind this approach and offer a definition of problematising categories as the practice of making conscious how socio-linguistic units of categorisation unconsciously organise our perception and can represent a problem for knowledge production. This practice, I argue, enables us to decentre ourselves from the taken-for-granted nature of those categories. Second, I develop a three-stage research method for problematising categories: noticing ‘critical junctures' when problematisation is called for, identifying the categorical problem through sensitising questions and reconstructing an alternative. Third, I demonstrate how problematising categories contributes to the research process by applying this method to my experience in problematising the binary pair ‘local’ versus ‘international’ in a research project on the environmental impact of Chinese investment in the Senegalese fishery sector. I show that problematising categories leads to more rigorous empirical findings and nuanced analysis in a way that is feasible within the frame of qualitative research projects. Overall, this article expands the practical tools for linguistic reflexivity and heeds the methodological call to make conscious and explicit choices for every dimension of our research.

Background

This article emerged out of the frustration I encountered as a junior researcher and teacher. On the one hand, I was theoretically and epistemologically grounded in discourse theory and critical qualitative research and thus working with the premise that language organises our perception and produces socio-political effects. On the other hand, due to the lack of dedicated pedagogical and methodological material, I struggled to tackle this problem in my own research practices. This absence also made me feel unequipped, as a teacher, to support students in their research journey. I needed tools to help them overcome some of the recurrent challenges they faced regarding reflexivity and language, be it trying to fit empirical material into categories that did not work or losing their self-esteem thinking that their cognitive capacity – rather than unproblematised categories and the lack of material to learn how to address them – was the problem. I wrote this article as a pragmatic and practical step forward towards addressing this issue.

Introduction

The growing interest in reflexivity in qualitative research has highlighted the ‘general lack of sufficient detail given over to the ‘how’ in relation to this process’ (Maxwell et al., 2020). To address this gap, a stimulating body of literature has emerged that aims to adapt methodological guidelines and tools to the practice of reflexivity. Some researchers have repurposed existing methods for the mission of reflexivity, such as visual methods (Darnhofer, 2018) and auto-ethnography (Woodley & Smith, 2020). Others have focused on formalising reflexive practices they developed to address their own reflexive needs. In the last 15 years, this literature has created a veritable toolbox for ‘reflexive research methods’ (Bryant & Livholts, 2007) that includes approaches such as ‘dialogical storytelling’ (Carter et al., 2014), ‘perspective taking’ (Finefter-Rosenbluh, 2017), ‘social identity map’ (Jacobson & Mustafa, 2019), ‘collecting sensorial litter’ (Hare, 2020) and ‘memory work’ (Bryant & Livholts, 2007). By making such reflexive practices explicit and sharing them with a large audience, this literature has successfully turned informal practices for reflexivity into ‘methodological things’ that can be debated, learnt and taught alongside other research practices.

However, despite qualitative researchers’ acknowledgement that language is not a neutral tool of communication but a social practice that affects knowledge production epistemologically and socio-politically, these initiatives have largely focused on positionality and situatedness at the expense of researchers’ use of language in the process of knowledge production. That is, these initiatives have focused on unpacking processes to help us become more aware of who speaks and from where we speak rather than how we speak. While the need to become more reflexive about our engagement with language is commonly raised in the literature – regarding, for example, research in cross-language settings (Bergen, 2018; Redman-MacLaren et al., 2019; Temple & Edwards, 2002) – practical tools to support researchers are largely missing. ¹

So far, qualitative researchers’ methodological interest in language has mainly taken two directions. On the one hand, the engagement with language as empirical material has led to abundant literature on methods of text and discourse analysis (Gee, 2011; Kuckartz, 2014; Schreier, 2012). On the other hand, attention to the words we use to produce our research has mainly focused on stages of the research design where researchers consciously engage with naming and defining the phenomena they investigate. Such activities, for example, include conceptualisation (Hupcey, 2002; Jabareen, 2009; Penrod, 2002) and the construction of categories in grounded theory (Giske & Artinian, 2007; Kelle, 2007; Scott, 2004). Yet, our routine use of language permeates every dimension of research outside these moments. Moreover, the implicit dimensions of language have implications outside cross-cultural research settings, as the socialisation to one’s language is largely an unconscious process (Schieffelin & Ochs, 1986). While scholars’ engagement with the conscious handling of language in research has established methodological debates and led to a plethora of practical tools, scholars’ routine uses of language outside these defined moments have not resulted in the same development. As a result, in the absence of adapted resources, researchers and students striving to become more reflexive about their use of language find themselves at a loss.

In this article, I argue for the need to develop methodological and pedagogical tools for linguistic reflexivity, following the reflexive initiatives already undertaken towards positionality and situatedness. I define language as the conventional system of communication through words, encompassing implicit and unconscious dimensions (such as connotations) that can be made conscious and explicit through dedicated methodologies. ² I define reflexivity as the practice of making conscious and explicit our practices, assumptions and dispositions; and the recursive use of these aspects of our socialisation as dimensions of our object of inquiry, in all phases of knowledge production. In doing so, ‘reflexivity leads us to recognise alternative ways of viewing ‘reality’ and prompts us to make explicit some of the world views which we and others bring to our research endeavour’ (Eakin et al., 1996, p. 158). As such, reflexivity gives rise to a ‘double knowledge’ – using the self as a resource to produce more analytically refined and ethical knowledge about the world and using the world as a mirroring object of enquiry to better understand the researcher’s self (Alejandro, 2018b, p. 191).

A methodology for linguistic reflexivity thus encompasses a set of practices that aims to structure and guide our reflexive work through a particular focus on our use of language. As a starting point to foster further conversation, this article develops a research method for problematising categories, a process often promoted as a promising approach to reflexively tackle scholars’ unconscious engagement with language (Alejandro, 2018a; Gillespie et al., 2012; Silvester & Topping, 2015; Sinha, 2006; Townsley, 2011) but that lacks formalised guidelines for its implementation.

To address this issue, I wrote this article with several objectives in mind. My first aim is to develop a pragmatic and practical approach that will help academics learn and teach how to problematise categories: a ‘formalised’ method to navigate categorical troubles that I wrote more as an analytical guideline than a rigid set of rules. My second aim is to demonstrate the analytical, empirical and ethical benefits of implementing ‘problematising categories’ to those who might not be familiar with this approach. This contributes to establishing the problematisation of categories as a core practice of qualitative research, as well as a specific reflexive strategy across different epistemological positions – be it to ensure that the implicit dimensions of our discourse match their explicit objectives or to investigate the perception and potential biases embedded within the categories we use when conducting research. My third aim is to foster an interdisciplinary discussion about the implicit practices that we engage towards linguistic reflexivity. This can help develop reflexive practices that focus on language use into ‘methodological things’ and make them more accessible to a large audience.

To do so, I engage in three successive endeavours. First, to help turn problematising categories into a ‘methodological thing’, I review the literature on the methodological problems that categories raise and introduce the traditions behind problematising categories. Here I show that the investigation of ‘categories’ and ‘problematisation’ both as social phenomena and as methodological processes provides a coherent conceptual framework for linguistic reflexivity. Second, I develop a three-stage research method for problematising categories that consists in (a) recognising the ‘critical junctures' at which problematisation is required, (b) identifying the potential categorical problem(s) through sensitising questions and (c) reconstructing alternatives. Finally, I demonstrate how the problematisation of categories fits in, and contributes to, the research process based on my experience in problematising the binary pair ‘local’ versus ‘international’ in a research project dealing with the environmental impact of Chinese investment in the fishery sector in Senegal, 2011. In this case study, I demonstrate how implementing linguistic reflexivity via the problematisation of categories helped me challenge preconceptions embedded in language that had led to a fieldwork crisis. This case study illustrates how problematising categories leads to more rigorous empirical findings and more nuanced analysis.

The Rationale Behind Problematising Categories

Researchers across the social science and humanities have put on the agenda the many issues resulting from the use of categories in knowledge production. They share their ‘distrust’ in categories such as ‘sex’ and ‘family’ (Azul, 2011, p.24) and question what binary categories such as ‘homo-hetero’ reflect and enable (Vicars & McKenna, 2015, p.420). They warn us against categories we may take for granted such as ‘community’ (Douglas & Gulari, 2015, p.360) and the homogenising effect of categories such as ‘youth’ (Tickle, 2017, p.66). They underline the ethical necessity to distance oneself from categories produced by the state and that reproduce its power structures – for example, ‘Scheduled Tribes’ in India (Tukdeo, 2018, p.184). They highlight processes of interpellation through which those who do not fit are ‘assigned to a difficult category and a category of difficulty’ (Ahmed, 2010, p. 582). They show how categories are socially produced, and how identification with categories comes with expectations of positionality and performance (Carbado, 2005). They put forward how processes of categorisation underpin processes of othering (Montenegro et al., 2017, p. 145) and entail practices of normalisation and exclusion, for example, when the category ‘gender’ was ‘either ignored in research or portrayed in devalued ways, leading to women being continuously framed as deviant in medicine and psychiatry’ (Mountian, 2017, p. 156).

To address such challenges, researchers mention ‘problematising categories’ as a solution. This approach to ‘category work’ (Ryen & Silverman, 2000) is commended as a means to better understand our object of scrutiny (Morgan, 2000, p. 147), avoid biases that obscure categories’ socially constructed nature (Pavlovskaya, 2002, p. 284) and shatter the illusion that categories – such as ‘homosexuality’, ‘lesbianism’ and ‘woman’ – have ‘coherent, unifying meanings’ across contexts and cultures (Blackwood & Wieringa, 1999, p.viii). In many ways, the problematisation of categories has become synonymous with good research, and the failure to do so when needed is reproved (Hart, 2006, p. 127).

However, it is unclear what scholars mean when they express their ‘commitment to problematizing categories’ (Maton, 2000, p. 155). What this practice entails, how to engage with it or where to start remains a mystery. As methodological and pedagogical resources demonstrating how to problematise categories in practice are missing, teaching and learning of this approach remain difficult, which also prevents the cumulative exchange of experiences and good practices between scholars around questions of categorical problematisation. As a result, and despite the imperatives to engage with such practice, we often struggle and waste time trying to make sense of some case before realising that the categorical dimensions of the words we use, rather than our analytical skills or our material, are the underlying cause of our concerns. As teachers, we observe that students are often stuck, trying to force their data or object of enquiry into categories of analysis that do not fit, rather than challenging these categories and the implicit connotations associated with them.

In this section, I review social science literature to understand what researchers mean when they talk about categories and the problematisation of categories and show how the analytical and empirical traditions behind this literature provide a rich grounding to develop ‘problematising categories’ as a methodological tool for linguistic reflexivity.

What Do We Mean by ‘Categories’?

Attention to the role of categories for cognition and perception started with the work of philosophers like Aristotle, John Stuart Mill and Emmanuel Kant. Kant, for example, established categories as ‘ontological predicates’ that enable us to think about objects by making a judgement about them via language Kant (1790: Introduction V). In that sense, categories refer to general properties of things (they do not inform us about individual characteristics of objects but characterise groups of objects with the same properties) and are mutually exclusive (for example, Kant highlights the category ‘possibility’ as opposed to ‘impossibility’, and the categories ‘existence’ as opposed to ‘non-existence’).

At the beginning of the 20^th century, the question of categories moved from philosophical to empirical grounds with its introduction to social sciences at the birth of these disciplines. Against the dominant Kantian interpretation that categories are derived from pure reason, Durkheim and Mauss (1903) argue that categories are comprised within ‘classificatory systems’ that are both acquired socially and productive of social order. They posit that categories organise our perception of the world not only by grouping objects that share the same predicate (e.g. both the writer and readers of this article can be qualified as ‘human’) but also through the relations established between categories. These relations can be oppositional (e.g. ‘human’ vs. ‘machine’ or ‘human’ vs. ‘animal’) or hierarchical when a category is comprised within another category and acts like a subcategory (e.g. the categories ‘men’, ‘women’ or ‘children’ can be comprised within the category ‘human’). Furthermore, they argue that neither categories nor the classificatory systems comprising the ensemble of their relations are innate nor inherent to individual psychology. Indeed, these scholars argue that how categories are relationally organised is influenced by the type of relations existing within society (a phenomenon they name ‘sociocentrism’). Societies produce classificatory systems, whose organisation may vary between societies. For example, the equivalents of the categories ‘human’, ‘men’, ‘women’ and ‘children’ may not have the same predicative value across languages, and how these categories relate with each other – for example, whether ‘men’, ‘women’ and ‘children’ are comprised under the category ‘human’ – vary across history and social boundaries. To summarise their argument, categories are articulated in classificatory systems that organise the way we perceive the world (including the socio-political world), and these systems are influenced by socio-political factors, and therefore may vary.

These early works paved the way towards a sub-field cutting across sociology, anthropology and linguistics that aimed to empirically explore the social functions and effects of categories in different contexts (Chave-Dartoen et al., 2012, pp. 93–100). Taking the case of the classificatory categories used by statistical economics in France, Boltanski (1970) questions sociology’s common distinction between ‘scientific’ and ‘popular’ classificatory systems; the former designating classificatory systems consciously constructed via scientific methodology while the latter designates systems unconsciously inherited. More specifically, he shows how this distinction has obscured the fact that most classificatory systems used by researchers are inherited rather than consciously constructed and therefore operate as ‘popular’ systems for the researchers who use them. Regarding pedagogy, Bourdieu and Saint Martin (1975) show how the categories used by Parisian high school teachers to assess their students vary according to the teachers’ knowledge of students’ socio-economic backgrounds.

These works illustrate how researchers and teachers may not be reflexive about the socio-political origins and effects of the categories they use. They emphasise that academic and non-academic use of language cannot be separated and that our positionality and situatedness as researchers and teachers are embedded in our everyday use of language. Like other aspects of language, categories are acquired implicitly through socialisation – that is, the lifelong process through which agents acquire the norms, habits and skills necessary to be part of society; and the way they organise our perception is largely unconscious to us unless we make an intentional effort to bring them to consciousness.

To sum up, categories must be understood as systems that both enable us to make sense of the world through language (a major aim of social research) and represent potential blinders for knowledge production due to the pre-organised and unconsciously inherited nature of their acquisition (which represents a challenge for social researchers). Empirical research on categories, therefore, sheds light on the methodological, epistemological and ethical implications of categories for social research. As a result, scholars emphasise the need to interrogate the definitional power of the categories we use by adopting an ‘epistemological vigilance’ that systematically engages in a lexicological criticism of common language' (Bourdieu et al., 1973). This methodological imperative dates back to Durkheim (1894/2009) himself who argued for careful reflexive work on common words – which we ‘dangerously’ inherit and which conspire ‘to make us see in them the true social reality’ – as the starting point of any scientific approach.

Problematisation as a solution to the problem of categories

The idea of problematisation of categories arose as a practical solution to the methodological-epistemological problem of categories. While agents commonly problematise outside of academic research (Fabre, 2006), the practice of problematisation seems particularly crucial to scholarly work since the ‘formulation of problems’ is the starting point of scientific enquiry (Popper, 1985). Accordingly, problematisation represents both the process through which social agents turn previously naturalised social issues into socio-political problems (Foucault, 1998) and a research practice. As a research practice, it aims to identify unthought problems within taken-for-granted knowledge or discourse in order to make them conscious and explicit so they can be addressed. As such, problematisation aligns with different epistemological paradigms that either approach reflexivity as an emancipatory endeavour or a way to identify unconscious biases.

External events, such as major societal crises that disrupt the status quo can help make conscious what has become taken for granted: a process commonly referred to as ‘denaturalisation’. For example, World War I’s reorganisation of Western European labour, which required women to take up responsibilities previously framed as ‘masculine’, is often considered as a catalyst for the denaturalisation of pre-war representations of ‘women’ and their alleged natural capacities (or lack of thereof) in comparison to ‘men’.

Rather than relying solely on external events, scholars have looked for practices that can be implemented at the level of the individual researcher to denaturalise linguistic habits and bring them under the scope of our critical scrutiny. Problematisation is one such practice: it can be used to shed light on the analytical blinders of the categories we have inherited, which otherwise unconsciously orient the type of knowledge we produce and constrain our analysis by forcing the ways we think and speak into pre-established rigid systems of categorical relations. Following the example developed above, using ‘boys’ and ‘girls’ as subcategories of ‘men’ and ‘women’ rather than using ‘children’ as a subcategory of ‘human’ may result in different research designs that can discriminatively gender the framing of children’s rights and (re)produce inequalities.

Interestingly, scholars have used problematisation not only to tackle taken for granted knowledge and discourses about the world but also the academic tools and practices they have routinised, for example, commonly used fieldwork methods (Fillieule & Pudal, 2010). Problematisation itself is but one approach that has been developed to challenge and question the naturalised use of language alongside other approaches such as deconstruction (which aims to disrupt the simplifying and definitional power of language) or discourse analysis (which empirically investigates the socio-political role of discourses) (Caputo & Derrida, 1997; Taylor, 2013). However, despite successfully demonstrating the need to pay attention to language, such initiatives have mainly developed practical tools in relation to other people’s use of language rather than our own use of language within the scope of a research project. As a result, guidance about how to problematise categories in practice in a way that is accessible for a large audience is lacking.

To conclude, interest in categories does not equate to the adoption of research practices that enable us to become more reflexive about our use of categories: for instance, Aristotle’s pro-slavery, sexist, inegalitarian, classist and racist posture illustrates that an engagement with categories is not per se emancipatory. Problematisation has been identified as a promising practical route to address this reflexive tension. I define problematisation as a process through which we identify and make conscious problems that exist but that have not been identified so far because they hide within dimensions of our experience that we take for granted. Rather than ‘creating problems where there are none’, problematisation aims to make conscious problems that already exist for us so that we can address them. Here, the unsuspected issues that we are looking for lie within the way we frame and represent our objects of enquiry via the linguistic categories we use. I define categories as socio-linguistic units acquired and shared through socialisation, which organise our perception not only through the way they make us identify some objects as similar and others as different, but also through the relations they establish between such groupings. Accordingly problematising categories represents the practice of making conscious how socio-linguistic units of categorisation unconsciously organise our perception and can represent a problem for knowledge production, in a way that enables us to decentre ourselves from their taken-grantedness. In that sense, problematising categories represents a promising approach for the practical implementation of linguistic reflexivity.

A Method for Problematising Categories

Considering the naturalising effects of language and the potential infinite regress of linguistic reflexivity, problematising categories can be challenging, especially when we attempt to do so without guidance. In this section, I present a three-stage approach to problematising categories:

- First, we notice critical junctures – that I define as key moments for when to problematise categories – to target which categories to problematise in priority;

- Second, we identify the nature of the categorical problem and how this problem may unconsciously organise our perception in a way that constrains the research process;

- Third, we expand our imagination beyond the naturalising effects of the categories we have unconsciously acquired and start reconstructing an alternative that takes as a starting point anticipated scenarios listed.

Through this process, I aim to provide a handrail for problematising categories that sensitises researchers and students for things to look at and promote linguistic reflexivity as a routine research practice.

Noticing the ‘Critical Junctures’

As we do not have time and resources to problematise all the categories we use, pinpointing key moments for when to problematise categories helps us focus our problematising initiatives to maximise their outcome. I define such moments as critical junctures to emphasise the weight of the potential reflexive momentum they can generate. Critical junctures require us to claim responsibility as socio-discursive agents engaged in a research activity when faced with the following alternative: either breaking from linguistic inertia by harnessing the tension that arises from the messiness of the world, scratching under the rigid surface of our categorical systems or walking away from shaking our linguistic habits in yet another missed opportunity to push research boundaries.

Examples of such critical junctures involve:

- when we are expected to make a methodological choice involving categories (e.g. the construction of an observation grid or sampling strategy; say, ‘youth Londoners’ regarding the implications of the categories ‘youth’ and ‘Londoners’);

- when we construct a research design whose ambition is to produce categorisation (e.g. developing a taxonomic system for text research, constructing a typology or choosing reference categories);

- when we conceptualise;

- when the introduction of new data does not fit our emerging analysis (see the illustrative case in the following section);

- when ethical issues arise from our engagement with the actors we study or when these actors seem uncomfortable with the way we express ourselves (see Dosekun (2015));

- when despite the explicit justification of our case, we intuitively feel that there is something that ‘does not fit’ or that our analysis does not quite do it justice;

- more broadly, anytime we feel stuck at any stage of our research process.

Here, one can note that problematising categories does not merely aim to help researchers become more linguistically reflexive outside the stages of research design where they consciously engage in language work (such as conceptualisation), but also within and alongside these moments. Indeed, the literature on categories explored in the previous section invites us to consider that researchers do not engage such conscious processes of language work from a blank slate perspective but rather come to it with pre-acquired categories that they need to become reflexively aware of. Problematising categories alongside such approaches supports researchers in reflexively exploring the implicit categorisation that may unconsciously influence these processes.

Identifying the Categorical Problem

Once we have identified that we are at a critical juncture, the second stage consists in asking oneself the question at the core of the method:

‘Could the categories I routinely use to conduct my research represent a problem that I am not aware of and that I should therefore identify as such so that I can address it?’

The question is simple. Yet, based on my personal supervising and teaching experience, as well as on my own struggles as a researcher, this question is neither taught as a core academic routine nor as part of the ‘first-aid kit response’ that comes to mind when we are stuck in our research. On the contrary, unproblematised categories that happen to be problematic are often experienced as personal struggles (the incapacity of the individual researcher to make sense of their cases) or denied (by forcing the empirical material to fit into the inherited categories rather than challenging the categories to better explain the material).

To guide researchers in answering this question, I developed ‘sensitising questions’ (Gillespie & Cornish, 2014) that aim to help readers identify common categorical problems they are likely to encounter in their research:

- Are the relations routinely established between categories relevant for my case study, or does this categorical relationship contradict elements of my case? For example, people who have experienced sexual abuse may be framed as ‘victims’ and be included under the category ‘vulnerable groups’ when it comes to social research ethical guidelines, while the use of such categories has been problematised on ethical grounds by members of this group, who may consider these categories disempowering and therefore opt for other categories such as ‘survivors’;

- What would the effect be of not using these categories to speak about a topic for which other scholars usually use these categories? Do these categories have heuristic value for my specific project or do I use them merely as a habit of language? (see the illustrative case in the following section);

- Is there a relation between the categories under scrutiny that polarises them as mutually exclusive? Can the characteristics referred to by the categories not happen simultaneously? If so, would conceptualising this phenomenon as a spectrum take away or add to the analysis (for example, see the engagement with the categories ‘able’ vs. ‘disable’ by De Schauwer et al. (2017))?

It is important to underline the paradoxical situation that the categories most in need of problematisation are those that may seem the least problematic for us, as the fact that there are perceived as non-problematic is often indicative of how thoroughly naturalised they are. Moreover, which categories should be problematised in priority varies between domains and researchers themselves as the naturalisation of categories is not universal and varies according to one’s socialisation. Accordingly, using the literature to identify which categories other scholars have already flagged as problematic can represent a promising starting point to problematise categories outside the scope of our imagination.

Reconstructing an Alternative

After having identified a problem dealing with categories, we need to expand our imagination of the options available to us and take the first steps towards reconstructing an alternative. Listing a few scenarios that we can anticipate helps us navigate the feeling that there is no other way of speaking the world than what the categories we use enable us to. One may consider:

- using the same categories but redefining them. Here, while we identify that implicit elements associated with the categories are problematic, we perceive that these categories have the potential to make a fruitful contribution to our research. In this case, it might be relevant to be transparent with the reader about our departure from the common connotations associated with the categories and how we are using them with a meaning we have more consciously crafted;

- using the same categories but reorganising their implicit relation. For example, scholarship developed the concept of ‘glocalisation’ (Robertson, 1994) to address the fact that in many cases the reification of ‘local’ and ‘global’ as two separated spaces produced by their separation in two distinct categories did not match the phenomena they aimed to study;

- substituting these categories by other categories. Categorical options available to us vary according to one’s own socialisation, for example, different people can be socialised to perceive the same social group as either a ‘terrorist organisation’, a ‘nationalist movement’ or ‘freedom fighters’. Exploring literature outside of what we are familiar with and actively looking for alternative categorisations might represent promising routes to expand the range of options from which we can consciously pick from;

- dropping these categories altogether. Not using the words everyone we know is using to describe a phenomenon may seem dauting. I picked this situation as illustrative case study in this article as it seemed the most challenging scenario (see next section).

Through these options, we see that problematising categories can be combined with approaches that engage language in qualitative research, such as conceptualisation and the development of coding frames. Besides representing a research method to be used on its own, problematising categories can therefore complement traditional analytical and linguistic work to foster scholars’ reflexive engagement. As with other types of reflexive endeavour, current norms allow researchers to decide how much of the problematising category work they want to share with their readership, which in turn encourages researchers to selectively share the information that supports their argument and demonstration.

In a sense, most challenges raised by language in academic research fall under the scope of the two following questions: Are the words I use the best linguistic options to support me in constructing, conducting and communicating my research? Do the words I use enable me to produce a knowledge capable of challenging the socio-political phenomena I aim to address, or do they carry implicit meanings that can potentially counter-perform these objectives and produce unwanted socio-political effects? I argue that problematising categories is a process that enables to answer such questions and that this process can itself be synthesised by asking oneself the following: could the categories I routinely use to conduct my research represent a problem that I am not aware of and that I should therefore identify as such so that I can address it? Rather than a mechanical procedure, I developed problematising categories as a three-stage method for linguistic reflexivity – which comprises focussing on critical junctures, identifying categorical problems and reconstructing alternatives – to help researchers better internalise routine strategies regarding their use of language.

Problematising Categories: a Demonstration

In this section, I illustrate the method for problematising categories by applying it to a research project that focuses on the environmental impact of Chinese investments in the fishing sector in Senegal. More specifically, I show that this method a) enhanced my reflexive awareness about how my unproblematised use of the categorical pair ‘local’ versus ‘international’ negatively affected my research and b) guided me in addressing this issue. To accompany readers through my problematisation process, I structure the demonstration via the three stages highlighted above, to show how unproblematised categories often emerge as tensions that we can resolve and overcome through problematisation.

In International Relations (the discipline within which I developed this project), the category ‘local’ emerged alongside other categories such as ‘transnational’ or ‘global’ to problematise the category at the foundation of the discipline: ‘international’. Literature within and outside International Relations has put forward the problematic character of the category ‘local’ and its pairing with categories such as ‘global’ or ‘international’ (Anderl, 2016, 2020), for example, in the field of peacebuilding (Mac Ginty & Richmond, 2013). As we will see, the existence of such warnings did not prevent me from running into categorical trouble nor did it replace the need to engage in problematisation in my own research project. More generally, taking a categorical pair as an example addresses the methodological concerns raised by qualitative researchers in regard to binary categories and their reifying effect (Parameswaran, 2001; Santos, 2014, pp. 45–46).

In this case, the critical juncture for problematisation arose during data collection when the introduction of new data shattered my preliminary results and led to a fieldwork crisis. The categorical problem I identified was that I was using the categories ‘local’ and ‘international’ as a habit of language and imposing them on a case to which they did not fit. Regarding this project, the best alternative was to drop these categories altogether and rethink my research design without them. During the demonstration, I write ‘local’ and ‘international’ in italics to help readers perceive how the unreflexive use of these categories influenced my thought process and shaped my understanding of the case until I problematised them.

Project Background

As part of a collective project about the environmental impact of Chinese investment in Africa, I was tasked with investigating the case of fishing in Senegal. At the time, fishing was a key sector for the country (employing 17% of the workforce and producing 30% of the export balance value (OECD, 2008; USAID & WestAfricaTradeHub/Dakar, 2008)). Understandably, the signing of international contracts – through which the Senegalese government gave fishing licences to foreign boats that potentially favoured an international, undeclared and unregulated fishing industry – raised local concerns regarding the diminishing of local fish stocks. In this context, studying the environmental impact of Chinese investment in the Senegalese fishing sector appeared as an interesting counterpoint to the China-in-Africa literature that focused mainly on energy resources.

The literature agreed that Chinese international activities in Africa had a distinctively harmful environmental impact on a vast array of local sectors but neglected the role of national and local actors, and regulatory contexts, in environmental issues (Compagnon & Alejandro, 2013). To assess this role in the case of Senegal, I conducted 22 interviews in July–August 2011 with fishermen, NGO representatives and executive officers working in governmental fishing and environment agencies in Dakar. I also conducted observations of the Port of Dakar and analysed online and archival governmental and NGO documents.

Preliminary Results

The interviews with the local actors (fishermen and NGO personnel) confirmed the international literature: everybody cheats when it comes to fishing regulations, but Chinese fishing practices are considered the worst. This consensus between local voices and international expertise sounded at first like a job well done.

Critical Juncture

When I went to the Directorate of Fisheries to collect data about the declared cumulative catch per country (to put Chinese boats' relative performance into perspective), I was shocked to learn that no Chinese boat was officially licensed to fish in Senegalese waters. Why did everybody agree about harmful Chinese fishing practices if there was no Chinese fishing in Senegal? The introduction of new data challenged my preliminary results and left me bewildered. This seemed like an opportune time to try problematising some of my categories.

Identifying the Categorical Problem

Could the categories I routinely use to conduct my research represent a problem that I was not aware of and that I should therefore formulate as such in order to address it?

As an answer to this question, I identified the categorical pair ‘local’ and ‘international’ as an unconscious problem that led me to make several methodological mistakes. More specifically, the implicit dimensions carried within these categories, which I was not conscious of, made me uncritically accept the consensus emerging from the interviews (and literature) at face value. I internalised the categorical distinction between ‘local’ and ‘international’ as characterising objects and people with inherently distinguishable attributes, which resulted in three types of blinders that affected my judgement:

- I essentialised interviewees as ‘local’ actors and the academic literature as ‘international’ scholarship;

- I uncritically deduced from this that so-called ‘local’ actors had ‘local’ voices while so-called ‘international’ actors produced ‘international’ discourses;

- I polarised alleged ‘local’ and ‘international’ voices and discourses. Because I romanticised ‘local’ voices as capable of producing more ‘authentic’ discourses than ‘international’ ones (which I unconsciously endowed with a potentially harmful or biased agency), I did not feel the need to critically assess their factuality or potential prejudices with the same scrutiny. This perception, implicitly entangled within my unproblematised use of the categories ‘local’ and ‘international’, made me bypass a basic rule of the methodology of interviews: the data we collect via interviews are less data about the world than data about the discourses and perceptions of the interviewees.

As we can note, these biases may not be specific to me or this project. There remains a common (while in some contexts problematised) assumption that the experiences of ‘local’ actors, expressed through their alleged ‘local’ knowledge and voices, ought to be different from those of ‘international’ actors. This assumption reifies social groups and denies the potential transnationalisation of discourse and cognitive frameworks – a circulation that can also occur between academia and the actors we study – while assuming that the circulation of discourses and prejudices is constrained by the imaginary geographical and sociological boundaries produced by the categorisation ‘local’ versus ‘international’.

To put it simply, I projected the categories of ‘local’ and ‘international’ onto the case without empirical information about the actual spatialisation of these discourses. As a result of my reification of the categories ‘local’ versus ‘international’, I reified the discourses I was analysing as if they were produced by distinguishable groups of actors operating in different social spaces (alleged ‘local’ actors of the ‘local’ contexts vs. alleged ‘international’ voices). I identified discourses produced by actors identified as ‘international’ as ‘international discourses’; I identified discourses produced by actors identified as ‘local’ as ‘local discourses’. This unempirical distinction gave the illusion that there was a form of triangulation in my information where data from different ‘sites’ confirmed one another: I mistakenly interpreted the ‘agreement’ between the allegedly ‘international’ discourses of the literature and the ‘local’ discourses of the interviewees as a sign that these discourses were empirically valid.

This acknowledgement, resulting from my problematisation of the categories ‘local’ and ‘international’, opened a new cognitive space within which to decentre myself from my linguistic socialisation. From there, I could then ask myself the following question: What would the effect be of conducting my research project without using the categories ‘local’ and ‘international’?

Producing Alternative Results

I decided to reconstruct my research project without using these categories to see if this initiative would help me overcome the biases I identified. As a result, I created three hypotheses that could explain the discrepancy between the discourses I collected and the official information from the Directorate of Fisheries: 1) the Senegalese administration does not have the capacity to monitor its coasts; 2) it turns a blind eye due to corruption; 3) both my interviewees and the literature share an anti-Chinese bias and there are no Chinese boats fishing in Senegal. The conclusion was that it was actually a combination of all three hypotheses and that the implicit blinders and values associated with the categories ‘local’ and ‘international’ had been preventing me from constructing a research design capable of grasping the complexity of this situation.

Confirming the first hypothesis, the government executives that I interviewed described the Senegalese fleet in charge of monitoring the coasts as insufficient. But this shortfall did not prevent the coast from being monitored. Foreign forces patrolled Senegalese waters in search of narcotrafficking boats and cooperated with Senegalese marines with issues of illegal fishing, such as the US army African Maritime Law Enforcement Partnership.

However, more hidden forms of undeclared fishing – such as those supported by corruption – fell outside the scope of foreign military patrols (in line with the second hypothesis). The state company China National Fisheries Corporation owned 49% of the shares of two Senegalese fishing and transformation companies, Sénégal Pêche and Sénégal Armement. These two joint-venture companies had been accused of using a Senegalese facade to benefit from tax exemption and get away with bad practices (GreenpeaceAfrica, 2014). This occurred in the context of growing corruption concerns regarding the son of then Senegalese President, who oversaw the cooperation with China.

Still, the fact that the interviewees focused solely on denouncing Chinese practices, despite the presence of 29 non-Chinese industrial foreign boats fishing in Senegal at that time, required explanation. The discourses of the interviewees specifically targeted Chinese fishermen (‘Chinese are everywhere even if you can’t see them’, ‘they will take it all…’) while describing practices I observed among other foreign fishermen (for instance, the fact of not speaking Wolof nor French, nor mixing with local populations). Rather than constructing a fair comparison of Chinese and other foreign fishing practices, interviewees’ discourses reflected the broader discontent with the growing Chinese presence in retail and construction (mainly Taiwanese until 2003), and the normalisation of Sinophobic discourses in Dakar, in line with the third hypothesis.

Conclusion of the Case Study

In this example, the unreflexive use of the categories ‘local’ and ‘international’ not only prevented me from understanding important elements of my case but also directly embedded biases in my research that resulted in the invalid interpretation of the data. By rethinking my research project without these categories, I reconstructed an alternative research design capable of incorporating data previously excluded and attained more rigorous and nuanced results. Moreover, taking the initiative to problematise categories represented a way forward during a difficult psycho-emotional situation. It was the first time I was conducting fieldwork in Senegal and on the fishing sector, and I already felt vulnerable and lacked confidence before going to the Directorate of Fisheries. In this context, problematising categories represented a constructive move, more beneficial than jumping in the rabbit hole of imposter syndrome and more rigorous than throwing under the carpet the problems with my preliminary results raised by the introduction of new data.

As we see, answering my research question did not require me to use the categories ‘local’ and ‘international’ at all in the end. In other words, the process of reconstruction following the problematisation of categories did not require me to substitute problematic categories with other ones. The fact that certain words are commonly used to categorise certain actors and situations should not prevent us from questioning their usefulness, nor from imagining that the world cannot be spoken without them. In this case, the unproblematised use of ‘local’ versus ‘international’ led to a fieldwork crisis that required me to problematise this binary pairing during the process of data collection. In doing so, I demonstrated that problematising categories, as a process of linguistic reflexivity, can be used as a ‘value-adding analytical device’ beyond data analysis (Eakin & Gladstone, 2020).

Conclusion

This paper aimed to build upon and expand the literature developing approaches to implement reflexivity in practice by formalising a method for linguistic reflexivity: problematising categories. The categorising function of language is an epistemological and methodological challenge that researchers cannot escape. As we aim to produce and communicate knowledge through language, the words we use act as categories that represent our basic – but double-edged – analytical tools. Categories enable us to organise our perception (e.g. by grouping as similar and distinguishing as different the objects of our attention) and thereby create perceptive blinders that we unconsciously acquire and potentially contribute to the (re)production of the socio-political order. While researchers have long highlighted the analytical and ethical challenges resulting from such processes of categorisation and our unreflexive engagement with language more broadly, they have not formalised research methods to tackle this issue. As a result, academics often struggle with or dismiss this challenge due to the lack of adapted tools.

To address this issue, I provided a conceptual framework and methodological guidelines to turn problematising categories into a transparent and accessible method for reflexivity. The method I developed comprises three stages. First, we identify critical junctures that I define as key moments for when to problematise categories. Second, we identify the categorical problem, for which I put forward sensitising questions to support researchers in their reflexive endeavours. Third, we imagine and start reconstructing an alternative. This third stage takes as a starting point the anticipated scenarios I listed. To show how problematising categories benefits qualitative research, I demonstrated how I problematised the categories ‘local’ and ‘international’ in a project focussing on the environmental impact of Chinese investment in fishing in Dakar, Senegal. By doing so, I aimed to encourage scholars to transform moments of doubt and struggle into reflexive opportunities for knowledge production and social change based on the questioning of the implicit and relational dimensions of the words they use.

It is likely that the few sensitising questions I outlined will not be the exact questions that will hit the mark for every specific research project. Similarly, binary categories are only but one example of situations experienced by researchers who might struggle with categorical continuums (e.g. more or less democratic) or categorical systems larger than a binary. I hope this piece will foster work that helps develop more practical tools for problematisation and the reflexive understanding of the role of categories in knowledge production, as well as instigate more reflexive explorations about the practice of problematisation itself as a socio-political and activist tool for qualitative research. Many traditions exist that have already engaged such reflections from discourse theory, to critical qualitative research, conceptualisation and analytical strategies. I am confident that building the bridge between these traditions and the existing pedagogical material on practical tools for reflexivity will enable such tools to become accessible to a broader audience.