Linguistic Characteristics of Texts: Methodological Notes on a Missed Step in Critical Discourse Analysis

Abstract

Critical discourse analysis is a set of theoretical and methodological devices used to analyze and challenge how we construct reality by looking for meaning behind words. The process of conducting critical discourse analysis is complex and, in the field of nursing research, is often carried out without regarding the different dimensions entwined in discourse. In this vein, the dimension that concerns the linguistic characteristics of texts is a highly informative one, but all too often left out of study results by researchers in nursing. This article aims to present some methodological notes of our experience conducting an analysis of linguistic characteristics within a critical discourse analysis doctoral research. We discuss the theoretical and methodological reasons why these characteristics should be widely recognized as a fundamental dimension in the framework of critical discourse analysis. We propose general recommendations to make this dimension of analysis more easily accessible and encourage other critical researchers to include the analysis of linguistic characteristics of texts in their research.

Introduction

Discourse analysis refers to a set of theoretical and methodological devices for the analysis of texts, whether written or spoken, to find meaning beyond words (Wodak & Meyer, 2001). There are different approaches that can be grouped under discourse analysis. One of these approaches is critical discourse analysis (CDA).

CDA is a method, but also a methodology, which positions itself both ontologically and epistemologically when addressing discourse. This critical and interdisciplinary approach to discourse sees language as a social practice, or as a means in the social construction of reality both constituted and conditioned by social practices (Blommaert & Bulcaen, 2000; Fairclough & Wodak, 1997). The connection between what is linguistic and what is social defines this particular approach to discourse best (Fairclough, 1992).

CDA allows an analysis of how discourse is shaped at a micro-level (i.e., through linguistic characteristics of texts), and how this micro level is connected via a meso-level (i.e., discursive practices) to a macro-level (i.e., social practices embedded in social, cultural, political, and ideological contexts in which language is used). Smith (2007, p. 61) states in this sense that “CDA thus works to mediate between pure linguistic analysis such as that of conversation analysis on one side, and poststructuralist studies of discourse on the other”.

Researchers who use CDA focus chiefly on the study of the way in which relationships of dominance, discrimination, power and control manifest themselves in the use of language (or in discourse) (Van Dijk, 2003). They are interested in understanding how these relationships, immersed in the beliefs, behaviors, actions and social practices that constitute social and political realities or orders, are unjust and detrimental to social welfare (Fairclough, 1995; Van Dijk, 2003). In this way, CDA shares the premises of Critical Theory in addressing the “social wrongs” that individuals face in today’s society (Fairclough, 1995).

Consequently, researchers who use CDA are also interested in exploring why social realities and orders are the way they are, what are the effects they have on contexts and what are the social mechanisms by which they are (re)produced. They seek out how these social realities or orders are considered and upheld as normal, naturalized, and institutionalized or, in contrast, are resisted or challenged through discourse discontinuities (Fairclough, 1995; Van Dijk, 2003). In this way, CDA deems power in social processes essential to the understanding of discourse (Fairclough, 2001). It is concerned with grasping how power creates and affects the operation of language, both to sustain the status quo and to challenge it towards social change (Fairclough, 2003; Wodak & Meyer, 2001).

Nevertheless, CDA does not consider language as powerful per se. “It is given power as a result of how it is used, who uses it, and the context in which it is used” (Wodak & Meyer, 2001, p. 10). The comprehensive study of who the subject that produces the text is, or to whom the legitimacy of using a certain language is attributed, and of the institutional context where discourse is produced and acquires its authority and its potential for application (Foucault, 1969), is what researchers using CDA focus on.

In the field of nursing research, the use of CDA as a research method and methodology is becoming ever more popular and is being widely used by increasingly prominent scientific production. Its popularity in this area, as opposed to how it has manifested itself in other areas where its development has a longer history, is accompanied by poorly described research practices that ensure its application in all its breadth (Smith-Merry, 2014). As a result, CDA continues to be poorly understood in nursing research (Cheek, 2004; Thorne, 2000).

Researchers in nursing often state that CDA they implement go beyond the linguistic characteristics of texts (Crowe, 2005; Smith, 2007). Thus, they pay limited attention to texts as units of analysis. They choose to focus on the thematic dimension of texts reporting directly on discursive and social issues emerging from them. This suggests that looking at linguistic characteristics ends up being marginal in most studies in nursing (Fairclough, 2003). Although analyzing and reporting linguistic characteristics of texts is fundamental in the scope of the interpretative and critical potential of CDA to capture the impact of discourse in the production of meaning and in the construction of the social world (Denham & Onwuegbuzie, 2013; Fairclough, 2003; Jorgensen & Phillips, 2002), both processes remain forgotten in a large part of CDA studies in nursing.

The objective of this article is to present some methodological notes of our experience conducting an analysis of linguistic characteristics within a CDA in a doctoral research in nursing. The article begins with the description of the research from which these methodological notes have been made. We then present the different stages of the data analysis process we conducted including illustrative quotes from the results of our research. From these insights, we discuss the theoretical and methodological reasons why the analysis and report of linguistic characteristics should be widely recognized as fundamental in CDA. We also develop general recommendations to make this level of analysis more easily accessible to nursing researchers. The article concludes by encouraging researchers in nursing who conduct their studies using CDA to include the analysis and report of linguistic characteristics of texts in their research to understand research phenomena more holistically.

The Study

The purpose of our study was to explore how nurses deploy their political agency to project their idea of social and political justice in public healthcare organizations and how they face the challenges and uncertainties of rethinking their institutional order when it does not resonate with their professional ethos. To explore this objective we used an ethnomethodological design (Garfinkel, 1991) supported by a theoretical framework that brought together the theoretical thinking of Nancy Fraser and Hartmut Rosa. This design reveals the nature of social orders through the study of how people try to make sense of the world around them from the analysis of genealogical relationships established between their social practices and the accounts told of them (Firth, 1995; Hassard, 1990; Lynch, 1993; ten Have, 2004).

This study was conducted in Mallorca (Spain). Participants in this research included central and peripheral voices (López-Deflory et al., 2022). Care nurses in clinical practice, nurses in middle management positions, nurse managers, and nurses in mixed positions (between care and management) within public healthcare organizations, as well as nurses in political, collegiate, or union positions were included as the central voices of the study. Physicians in clinical practice and managers of public healthcare organizations were included as the peripheral voices of the study. The inclusion of peripheral voices identified in the central voices’ accounts contributed to a better understanding of how central voices’ identities and the economic, political, social, and cultural context they inhabit are dialectically constructed (Musaeus, 2017).

Participants were recruited by intentional sampling (Koerber & McMichael, 2008) and segmented by sex according to the representativeness of women and men in each of the participants’ professions. This information was obtained from the data provided by the official professional associations. We included 31 central voices and 5 peripheral voices: 15 nurses in clinical practice (11 women, 4 men), 4 nurses in middle management positions (3 women, 1 man), 5 nurse managers (3 women, 2 men), 2 nurses in mixed positions (1 woman, 1 man), 3 nurses in political and/or collegiate positions (all participants in this profile were women), 2 nurses in union positions (1 woman, 1 man), 4 physicians (2 women, 2 men) and 1 healthcare organization manager (the participant in this profile was a man). All participants signed a written informed consent form to participate in the study.

Techniques of data collection included in-depth individual interviews, a sociodemographic questionnaire, and a field diary. The interviews were conducted in the participants’ preferred language (Spanish or Catalan) and audio-recorded to later undergo transcription and analysis. The sociodemographic questionnaire collected data related to sex, work experience, current scope of professional practice and job position of each of the participants. The field diary included written and audio-recorded descriptive, theoretical and methodological notes (Phillippi & Lauderdale, 2018).

The textual corpus of the study was composed of the transcription of the participants’ verbal and nonverbal communication. This stemmed from the integration of the field diary descriptive notes in the interview transcripts. Specifically, the participants’ nonverbal language was collected through the audio recordings, when we looked for paralinguistic elements, and through the field diary descriptive notes, when we looked for kinesics, proxemics and other elements at the specific moment of the interview in which they appeared. The transcription of the interviews and the integration of the field diary notes were carried out by the researcher who conducted them as a strategy of rigor in the research process.

We conducted CDA upon the data generated. We used Norman Fairclough’s approach (1992, 1995, 2001). Fairclough is considered one of the founders of CDA applied to sociolinguistics (Wodak & Meyer, 2001). Fairclough proposes a three-dimensional framework for analyzing discourse. The first level, namely discourse as text, consists of the analysis of how linguistic characteristics are articulated in texts to produce meaning. This level focuses on the study of vocabulary, grammar, cohesion, and structure of texts. The second level, namely discourse as discursive practice, consists of the analysis of the production, transmission, and consumption of texts. This level focuses on the study of strength, coherence, and intertextuality. The third level, namely discourse as social practice, consists of the analysis of the social context in which texts are produced. This level explores how power and ideologies manifest themselves in a given sociocultural context and how they affect the construction of identities and institutions within it (Fairclough, 1992, 1995, 2001; Fairclough & Wodak, 1997).

From our point of view, Fairclough’s theoretical and methodological contribution is one of the most holistic ones to guide the critical strand of discourse analysis. It works as a didactic-pedagogical methodological blueprint to guide CDA and capture the complexity of the meaning of discourse in all its dimensions. It is particularly useful for those researchers who are new to this method and who feel the need, as we have felt, to stop at the analysis of the textual level of the corpus elaborated from the data generated in the research process (Fairclough, 2003).

The Stages in the Process of Linguistic Characteristics Analysis in CDA

Data analysis in our study was continuous and transversal (Flick, 2014). The stages that made up this process did not take place one after the other. On the contrary, these occurred simultaneously and were intertwined. This forced us to go back again and again at the different stages to shape the results of an analysis process that was initiated simultaneously with the data collection and lasted until the end of the research.

During the analysis we often found the need to say more in detail about the linguistic characteristics of the participants’ accounts. We considered them highly informative data. This need was sometimes beyond what we felt able to report since there is uncertainty in the literature about how to analyze the data that make up the most purely textual dimension of discourse (Onwuegbuzie et al., 2010). To overcome this difficulty, we find it useful to structure the data analysis process into three stages: the preparation of the textual corpus, the organization of the data, and the dialogic approach to the results. These three stages need to be understood as interlocking stages even though they are presented linearly separate in this article.

In this article, we do not aim to present the results of our data analysis process. Instead, we sought to focus on the description of the methodological notes reporting how we have approached the analysis of linguistic characteristics within the CDA we conducted. The explanation of each of the stages mentioned above is as follows.

Stage 1: Preparing the Textual Corpus

The first stage of the data analysis process consisted in the preparation of the textual corpus to be analyzed. In our study, data was mainly garnered from interviews and researcher’s diary descriptive notes. Both had to be transcribed verbatim (Flick, 2014). According to Jorgensen and Phillips (2002), a transcription system considering the objective of the research has to be chosen. Some research purposes require a detailed microlinguistic analysis while others require a less detailed textual analysis. The sensitive nature of the topics addressed in our study, such as those related to the recognition of (in)justice experiences, and the researcher’s first impressions of the participants’ nonverbal language, such as the existence of certain patterns in the relationship between the participants’ verbal and nonverbal language, pointed out that the linguistic characteristics of the textual corpus would be relevant to data interpretation. Participants’ accounts revealed data of interest hidden behind these characteristics. We therefore chose to make transcriptions that would allow for a detailed microlinguistic analysis to be carried out later.

The researcher who conducted the interviews transcribed them immediately after their completion whenever possible. This allowed her to avoid forgetting important details regarding the connection between the participants’ verbal and nonverbal language. She transcribed both the verbal and nonverbal behavior of the participants. Following Gorden (1980), four major types of non-spoken expressions which the participants used to support their spoken expressions were transcribed. These included kinesics (i.e., body language including facial gestures, body movements and postures), proxemics (i.e., the organization of space in communication denoting the relationship between the interviewer and the interviewee), and chronemics (i.e., participants’ use of time) which were collected descriptively in the field diary, and paralinguistics (i.e., behaviors linked to vocal expression) which were captured through the audio recordings. At the end of this stage, we obtained a largely detailed textual corpus ready to be organized and then analyzed.

Stage 2: Organizing the Data

The second stage of the analysis process consisted in the organization of the data. To this end, we read each transcript several times to holistically understand its meaning. We then carefully read each transcript putting it in relation to the rest of the transcripts to address the three levels of CDA (the analysis of discourse as text, as discursive practice, and as social practice) by encoding the data.

The coding we carried out did not focus solely on what the participants said. It also focused on how the participants said what they said. The codes we developed thus included different codes that alluded to the thematic content and the formal content of the data. Both encodings were parallel and remained in continuous dialogue during the data analysis process.

We grouped the codes created into subcategories and subcategories into categories (Flick, 2014). We specifically formulated three different but interrelated categories. These categories dealt with (a) the way nurses thought about themselves and their profession (i.e., about their individual and collective professional identities and the meaning they attributed to being nurses); b) the acknowledgment of the forms of (in)justice nurses felt they faced in healthcare organizations (i.e., the inequities identified in the distribution of time, workload, material resources, spaces, and remuneration, as well as in their professional recognition and representation); and c) the practices nurses mobilized with an emancipatory yearning in the face of the forms of injustice identified (i.e., their emancipatory collective project, their ideas of change, and the specific actions performed so that the organization they inhabit could resonate with their professional ethos).

The in-depth analysis of the data that constituted the three categories made us aware of a feature. The codes that made up each category differed from each other regarding the interpretative goal they fulfilled in the analysis process. This difference led us to divide each category of analysis into three sections: a preliminary section, a central section, and an impact and transition section. We considered this way of reporting the study results the most suitable one to untangle their complexity and to highlight their dialogical relationship and the permeability between them.

The preliminary section considered how the participants talk about the general theme of each category. This section was characterized by focusing on the study of the linguistic characteristics of the textual corpus. The central section considered what the participants talked about. This section was characterized by being descriptive and by staying very close to the explicit data present in the text. The impact and transition section considered what consequences were triggered by the results included in the central section. This section was characterized by being more interpretative and by remaining somewhat distant from the explicitly generated data. This section explored the implicit effects of the results of the central section and worked as a background and link to the next category (and, in the case of the last one, to the discussion of the results).

The methodological notes on the preliminary sections we developed during the data analysis process are presented below. We do not address the other two sections because the specific objective of this article is to demonstrate the analysis of the linguistic characteristics of the textual corpus of our study and, as it has been mentioned, this analysis was conducted in the preliminary sections.

The Preliminary Section of Data Analysis

The preliminary section represented the first section of each category of data analysis. Its aim was to collect the results at a level of analysis focused on discourse as text. These results provided information on how the participants talk about the issues specifically included in a category, but also (albeit in a secondary way) to the issues included in the other two categories.

The codes that integrated the preliminary sections reflected the existence of patterns around how the participants (un)consciously mobilized discursive strategies with discursive effects to express themselves. In this sense, through their use of language the participants presented particular ways of (1) relating to themselves to introduce themselves to others as nurses, (2) communicating their experiences of (in)justice in healthcare organizations, and (3) (re)negotiating their status as nurses in front of others and in the world. These discursive strategies triggered effects that oscillated between (re)production and challenge to power relationships that make up healthcare organizations institutional status quo.

The aforementioned strategies were incorporated into an index of issues of micro-textual nature that had to be remembered each time the rest of the data was read and interpreted. Hence, the function of the mobilized discursive strategies analysis was two-fold. On the one hand, it ensured data contextualization. It offered a background for the thematic and discursive interpretation of the results. On the other hand, it guaranteed a data comprehensive approach. It provided a tool to capture the meaning of the results in all its complexity. We now present the main results included in the preliminary sections of the data analysis process we conducted.

The analysis of the linguistic characteristics of the results that made up the first category suggested that participants made twists in their voices when talking about themselves. These switches were manifested in the grammatical person, and the number, gender, time and space from which the participants talked about what it meant for themselves to be nurses. These features played an important role in understanding the construction of nurses’ individual and collective professional identities. We captured these issues by underlining certain information units and creating an index of the contexts of their use in the margins. This index allowed us to recognize patterns among the participants’ accounts and later explore the discursive effects that emerged from them.

We identified the personal pronouns and possessive pronouns in the textual corpus (Fairclough, 1992; Fromkin et al., 2011). The participants used to switch between the first and third grammatical person. They started talking about themselves in the first person (I) to introduce themselves to the researcher, but early on they changed to the third person (he/she) to project themselves as the nurses they aspired to be or thought they should be but were not due to internal and external limitations. Some participants’ narratives explicitly reported how nurses used this switch to hide their selves behind an ideal self.

I am a nurse and for me being a nurse is… (he pauses) Being a nurse is… (he looks upwards) Well… I have the ideal (he lowers his tone of voice) that a nurse is a professional who does everything for the good of the patient she cares for [Male nurse in a mixed position 1]

We identified the singular and plural grammatical persons in the textual corpus. At the beginning of the interview, the participants used to express themselves in singular as individuals who were and did on their own, while as the interview progressed, they used the plural as members of a wider community, with a common professional identity, who experienced the same forms of injustice. The context of use of this switch and the paralinguistic characteristics that accompanied the participants’ narratives suggested that nurses felt stronger when talking about themselves in the plural. They used the plural and a high and confident tone of voice to defend their individual experiences of injustice in front of others.

(She lowers her tone of voice) I have felt invisible lots of times. Nurses are invisible (exclaims)! (She uses a higher tone of voice) We do a lot of invisible work. If we don’t make our work visible, it’s s as if we don’t work, and if we don’t work, it’s as if they don’t have to worry about ourselves [Female nurse in clinical practice 1]

We also identified the masculine and feminine gender endings that the participants used to talk about themselves. As opposed to English, there are separate endings for gender in Spanish and Catalan. Regardless of their sex, the use of the term “nurse” in feminine, as is usual in both languages, was prominent because nurses represent a mostly female group, marked by historical-patriarchal roots, dedicated to caring for others. However, the participants used to speak of themselves in the masculine when they showed themselves as professionals with high technical skills, disciplined characters, responsibilities, and knowledge guided by a professional practice based on scientific evidence. Most of the participants reproduced social gender roles through language. In fact, some of them, those who held a more developed critical thinking in comparison to the others, explicitly reported being aware of this in their narratives.

I believe that a nurse (female ending) is a professional who accompanies people from birth to death. (…) We (female ending) focus a lot our professional discourse on care and care is very ethereal and there are many people who care, right? (…) We (feminine ending) must polish the discourse. (She uses a higher and more confident tone of voice) We as nurses (male ending) are increasingly trained and qualified professionals, we have a high degree of training, and we base our practice on evidence [Female nurse in a political position 1]

I always deliberately use the term “nurse” in masculine (…) because I believe that if you continue using “nurse” in feminine and “physician” in masculine what you are doing is perpetuating roles that are deeply rooted in society (…), you are perpetuating a hierarchical relationship symbolized in language [Female nurse manager 1]

We also identified the verb tense conjugations and the adverbs of time and place used by the participants. The past tense was commonplace in the nurses’ accounts. Adverbs such as “historically”, “before”, “formerly” and expressions such as “years ago” or “when we were technical healthcare assistants” (referring to the professional figure that, in Spain between 1953 and 1977, represented those who later became nurses, a figure whose functions were dependent on physicians, exclusively auxiliary and focused on technical practices (Miró et al., 2012)) were commonly used by the participants for reviewing their individual and collective history seeking reasons to understand why they are the way they are today. The thematic content that accompanied these linguistic characteristics was given continuity in the present tense with adverbs such as “still” or “yet”. The scarcity of the future tense in the participants’ accounts marked their meager longings and aspirations and their little hope for a better future.

I think we have to be aware of our history as a profession: where we come from, where we are and where we should be going. (He pauses) We cannot ignore where we come from because our historical position of subordination allows us to understand the reasons for many things today [Male nurse manager 2]

It is true that we have evolved with respect to our ancestors (…) But, when it comes to practice, we are still practically at the same point as before (she sighs). We don’t put on the physician’s gown, we don’t bring him coffee —thank God— as used to happen (she laughs), but we still don’t have a voice in politics or in management [Female nurse in clinical practice 1]

We nurses have made a lot of progress in a short time, but… (He pauses) We have been like this for so long (referring to being exposed to forms of injustice) that it is very difficult to change our situation (…) There are things that, however much we want, are not going to change (he grimaces) [Male nurse in clinical practice 3]

Finally, the data analysis suggested that the participants placed themselves “below” other professionals (mainly physicians) and some factors (medical orders, pressures, responsibilities, workloads). They also seemed to place themselves “in the middle of” healthcare organizations as links between other professionals and patients and families; “next to” patients, families and population; rarely “on top of” someone or something; “at the front of” the healthcare system; “behind” a large number of invisible processes; “close to” the area of care; and “far from” issues that were beyond their field of action, such as politics or research.

In short, in the first category of data, the participants assumed different positionalities as subjects. They manifested different I’s. They did not express themselves from a single voice. They adopted plural and dynamic voices. They showed different selves from which to speak and act according to the context and the intentionality pursued in each of them. They constructed multiple nursing selves through discursive strategies linked to linguistic characteristics in the text, specifically, through switches in the use of the first and third grammatical person, singular and plural, masculine and feminine, past, present and future, and spatial locations.

The analysis of the linguistic characteristics of the results that made up the second category suggested the existence of patterns in how nurses dealt with the forms of injustice they recognized in healthcare organizations. This description stood out with the use of the second grammatical person, the arrangement of the plot of their accounts, the tendency toward a rhetorical transformation of reality, and the recurrence of exemplification and comparison. We identified these issues by underlining the strategies that the participants deployed in their acts of speech and noted the contexts of their use in the margins. This index allowed us to acknowledge the existence of patterns among the participants’ accounts and later explore the discursive effects that emerged from them.

We identified the personal pronouns of the second grammatical person throughout the textual corpus. Data analysis suggested that nurses used to switch to the second grammatical person to dialog with themselves, to analyze their most sensitive, painful, or uncomfortable experiences from the outside, and to involve the researcher in their experiences.

Based on the staff available, my boss decides where he puts me (…) He does not consider (she pauses for a long time) neither my knowledge, nor my experience, nor my personal situation. Know what I mean? The fact that you are in a different operating room every day, that your preferences are not considered… (she winces) (…) He always puts me in otolaryngology, in ophthalmology, in things that I hate deeply, that I don’t like (she laughs). You never get a break (she sighs). A reinforcement is given to you the day there is an intervention that is very, very, very complicated, very complex. As they know you can deal with it… Then, that day (she emphasizes), they say: “No, X, you are a really good professional…” (she expresses mockingly) [Female nurse in clinical practice 1]

We analyzed the flow of ideas in the participants’ narratives. We perceived that nurses arranged their stories on two narrative levels. They began by discussing the most easily patent issues (their lack of time, overload of work, lack of physical space and resources to work properly, and poor salary). They expressed themselves about these topics with an energetic and confident tone of voice and built their sentences around verbal nuclei that denoted certainty. Sentences starting with “I know that” were recurrent in the participants’ accounts. They finished discussing the less easily evident issues (their professional recognition and representation). They expressed themselves about these topics with a tone of voice that denoted prudence and with a certain pre-critical repression, through prolonged silences before starting to talk about compromising issues or through expressions that sought to confirm that what they talked about would be kept confidential (such as, “I’m going to tell you something, but don’t let anyone else know”, “Nobody will hear this, right?“, “I don’t know if I should tell you… well, you’ll cut me off”), and post-critical repression, through expressions of minimization or regret of having said what they said (such as, “I don’t know if I should have told you this”). They also built their sentences around verbal nuclei that denoted doubt. Sentences headed by “I believe”, “I think” or “I suppose” were more common in this second narrative level. Furthermore, they usually communicated their most favorable and positive reflections first, leaving the most unpleasant ones for later separating them with adversative conjunctions, such as “but” or “nevertheless”.

(After reflecting on the most comforting issues of being a nurse, she switches to a hopeless tone of voice), but everything around us… (prolonged silence) Everything around the nurse and everything that has to do with how we are looked at or how we are cared for by organizations… (she sighs) That leaves much to be desired (she puts the accent on “that” by adopting a derogatory intonation) [Female nurse in clinical practice 2]

We also analyzed the use of rhetorical resources, summary, and humor in the participants’ accounts. Data analysis suggested that nurses used to transform their realities by using metaphors (especially linked to war) to protect themselves and trivialize the harshness of their working conditions. Expressions such as “nurses are fighters”, “our day to day is a struggle”, “we work in trenches”, “we are a threat to others”, “we have to gain ground”, “all for one”, “elbow your way through” or “conquer spaces” were found in all the participants’ narratives. These appeared in the framework of the report of what the status of nurses in healthcare organizations was like today.

We still have to demonstrate the skills we have (she sighs), but we still also have to prove them to ourselves because among ourselves we do not believe in them (she exclaims)! Yes… (Prolonged silence) I think it’s like a constant, constant, constant struggle (she hits the table) [Female nurse in middle management position 1]

In addition to metaphors, the participants also used other strategies to protect themselves. Such strategies included summary, observed in their tendency to leave their sentences incomplete, to change the subject of sentences quickly or to briefly describe their most negative experiences, and humor, observed in their recurrent ironic tone of voice and laughter.

Anyway, you get used to things (referring to her experiences of injustice) and get down to it (she laughs). What else did you ask me? [Female nurse in clinical practice 4]

We finally explored the resources participants used to justify their reflections. On one hand, data analysis revealed that nurses focused on describing anecdotes to share their experiences of injustice as it was difficult for them to give them a theoretical and global meaning. Expressions such as “I don’t know how I can explain it to you”, “I’m going to give you an example because if not, I don’t know how to explain this” or “I’ll give you an example and it will be easier to understand” were recurrent in their narratives. On the other hand, the participants used to compare themselves to others to better understand and make others understand their status in healthcare organizations.

If I look at myself and ask, “Am I well-paid?”, “Yes, I am” (…) But if I compare myself to others (he pulls a face)… (Prolonged silence) I think I’m getting too little for what I do (he complains) [Male nurse in clinical practice 4]

At the social level, the physician still has much more status. He is treated with more respect than we (nurses) are [Female nurse in clinical practice 5]

In short, in the second category of data, the participants undertook different strategies to report their working conditions. They did not communicate them arbitrarily. They communicated them in a particular and highly informative way. Nurses showed different strategies to be aware of the status they occupy in healthcare organizations. This status may have been informed in part by analyzing the linguistic characteristics of the textual corpus.

The analysis of the linguistic characteristics of the results that made up the third category suggested the existence of patterns in the way the participants coped with the report of how they behaved in the face of the forms of injustice they recognized in healthcare organizations. This report was characterized by the assumption of nurses as subjects committed to an emancipatory cause, the articulation of their speeches as if they were political speeches, and the creation of utopian scenarios in the construction of the professional scenario to which they aspired to. These resources played an important role in empirically understanding how nurses deployed their political agency in healthcare organizations. We identified these issues by underlining in the text indicators of nurses’ intentions towards their emancipation and creating an index of the contexts of their use in the margins. This index allowed us, as in the other two categories, to recognize patterns among the participants’ accounts and to explore the discursive effects that emerged from them.

We identified the particularities of the participants’ grammatical use of the first person in the last part of the interviews. Data analysis suggested that nurses tended to build their sentences from “we” or “nurses” to share their individual reflections from a collective perspective. The solidarity and sisterhood that emerged from the participants’ narratives at that time endowed their statements with a logic of ethical and political action from which the necessary courage was developed to take responsibility for undertaking practices aimed at nurses’ emancipation.

We are, and I am (he uses an explanatory tone of voice), healthcare professionals (…) (he opens his eyes and raises his index finger) Care is either covered by us or no one else bothers and, what’s more (he emphasizes), we must cover it! (he exclaims) [Male nurse in clinical practice 4]

We must be more demanding. We need a little more drive to assert ourselves and to raise our voices together and say: “Hey, we’re here”, you know? [Female nurse in clinical practice 5]

We must worry about ourselves, about the group, about taking care of each other, about seeing that if we need something we have to ask for it… (…) In the end, we are the ones who have to support each other (she shakes her head) [Female nurse in clinical practice 6]

We analyzed the meaning of the verbs participants used. Data analysis suggested that, in the last part of their interviews, nurses preferred verbs with performative meaning, such as “do”, “decide” or “can”. They replaced the verbs with reflexive meaning that were prevalent in the second category, such as “think”, “believe” or “suppose”. The participants thus assumed themselves as subjects who could not only “think” and “say”, but also as subjects who could also “make” changes happen in healthcare organizations.

We also studied the mood of the verbs participants used. Conjugated verbs in the subjunctive (verb form that expresses something desired or imagined) from which nurses expressed the conditionality of the ideas of change they proposed to transform healthcare organizations, and in the imperative (verb form used to give an order) from which nurses expressed the orders whose fulfillment would lead to those transformations were dominant.

I wish I had enough time to take good care of patients (she sighs) [Female nurse in clinical practice 7]

(He uses a confident tone of voice) We must believe that we are as qualified as other professionals and we have to be willing to participate in teams by contributing what only we, as nurses, can contribute: our perspective, our professional outlook. We must do it well, but above all, we have to dare to do it because we are used to remaining silent [Male nurse manager 2]

We explored the flow of ideas of the last part of the participants’ accounts. Data analysis suggested that there were patterns in how nurses articulated their speeches and in how the nonverbal language manifested itself within them at that moment. Specifically, they chose to explore the impacts of their working conditions, evaluated their significance, predicted their consequences, and elaborated the practices that should be undertaken to remedy them. They left little room for silences and pauses. They expressed themselves with a confident and higher tone of voice, and with firm body postures. They supported their ideas in the use of arguments, especially authoritative and probabilistic. They tried to convince the researcher of what they were proposing while they constructed themselves as legitimate subjects of a revalued condition of status in healthcare organizations.

I’m convinced that the only way to change all this (referring to the forms of injustice she previously recognized) is for people to know what a nurse is and does, for people to know that we are the center, that we are the point of union, the intermediaries between many collectives (…), that having more nurses makes the quality of care better. (She adopts a body posture that exudes confidence) Look, for there to be, for example, fewer nosocomial infections, it has been shown that what it is needed is more nurses, not better antiseptics. What it is needed is more nurses who can leave a room, who can wash their hands properly, and not go “boom, boom, boom”, fast! (she exclaims banging on the table) [Female nurse in clinical practice 2]

We finally identified textual indicators of a pessimistic attitude regarding the real possibilities for organizational changes that could transform healthcare settings into fairer environments. In this sense, at the end of their interviews, nurses used to regain a low tone of voice, an upright body posture and issued expressions that displayed the utopian construction of their aspirations.

Now I’ll introduce you to something very utopian (she expresses with a lower tone of voice) [Female nurse in a collegiate position 1]

This is utopia; it is utopia! (she laughs and shakes her head) [Female nurse in a union position 1]

All this should not be utopia, but something that could be reality (she complains and drops her shoulders) [Female nurse manager 2]

In short, in the third category of data, the participants constructed themselves as agents rather than as simple observers of what happened to them in healthcare organizations. They embodied subjects capable of undertaking practices to transform healthcare organizations into environments considered fairer and more resonant with their professional ethos. They deployed their political agency. This result could be understood in part through the analysis of the linguistic characteristics of the textual corpus.

In sum, the data preliminary analysis allowed us to better understand the complexity of nurses’ identities, the variability in the strategies through which nurses faced their working conditions, and the multidimensional deployment of their political agency inside and outside the text.

Stage 3: Establishing A Dialogue Between the Sections of Data Analysis

The preliminary sections of each category of analysis, that is, the study of how the participants talked about the issues included in each category, allowed us to gather the necessary pieces of the puzzle to begin to look at the bigger picture of the data generated and conduct CDA in all its breadth. These preliminary sections were placed in constant dialogue with the central and the impact and transition sections of our analysis. In other words, how the nurses talked, what they talked about and what consequences resulting from how and what they talked about were analyzed simultaneously. In Fairclough’s terms (1992, 1995), the analysis of discourse as text converged with the analysis of discourse as discursive and social practice.

The analysis of discourse as text was conducted together with the analysis of discourse as discursive practice. Thus, while analyzing the linguistic characteristics of the textual corpus we wondered how the text shaped certain social practices and how the text was shaped in turn by those social practices. We also wondered how the discourse reflected the social order, what were the signs of its rupture and what were its main contextual consequences. As the analysis process progressed, we undertook the analysis of discourse as social practice. We wondered about the broader context in which discursive practices operated, the ideologies being reproduced or challenged, and their broader (political and social) consequences (Fairclough, 1992). The theoretical lenses on which our research was framed were essential in this last stage of CDA. We will not offer more details about this third stage of data analysis because it moves away from the objective of this article.

Discussion

This article sets forth methodological notes of our experience conducting CDA in a nursing research. We have recognized discourse analysis as text, that is, the analysis of the linguistic characteristics of texts, at the same level of importance as discourse analysis as discursive practice and as social practice. We hope to have shed light on the need for researchers in nursing conducting CDA to consider the linguistic characteristics of texts as highly informative data to capture both spoken and unspoken utterances, this is what is said and what is said without being said through texts.

With the premise that everything is data (Eakin & Gladstone, 2020) and that in language everything matters (Fontana & Frey, 2005), we consider that the missed step of analyzing and reporting linguistic characteristics of texts in nursing should be re-assessed if CDA is to be conducted from a genuinely critical ontological, epistemological and methodological position. The concept “critical” is so widely used that its potential to address and analyze what happens at a micro level is often forgotten.

The analysis of discourse as text in CDA should be considered as complementary to the analysis of discourse as discursive and social practice (Fairclough, 1992, 1995). In fact, leaving one of the three dimensions of CDA unaddressed jeopardizes the critical character of this strand of discourse analysis and the very definition of CDA as the study of language as an element of social practice (Fairclough, 2003; Van Dijk, 2003).

Analyses of textual corpora that solely focus on their thematic content invariably omit features of texts that have important implications for how discourse is understood. By also addressing their formal content, researchers in nursing can enrich knowledge production (Flick, 2014). We therefore recommend a parallel data codification of the thematic and formal content of texts. This will contribute to a better understanding of relationships between discourses and other elements of social life, these including social relations and power relations, ideologies, social institutions and organizations, and social identities (Fairclough & Fairclough, 2012; Van Dijk, 2003).

We offer below some general recommendations for researchers interested in applying the methodological approach of discourse analysis we have presented in this article. These general recommendations are specially aimed at supporting researchers whose expertise is not in linguistics. We do not intend these to be considered a guide, but general insights serving as inspiration, since there is no singular way to conduct CDA (Crowe, 2005).

We recommend researchers, whatever the data collection techniques used in their research, to elaborate a detailed textual corpus that later allows an exhaustive analysis of its thematic or formal content. Despite the scarcity of time faced by researchers, we insist on the importance of this first stage of analysis. The accuracy of textual corpora is decisive for the type of analytical questions that can then be asked to texts.

We recommend a preliminary data analysis. A preliminary analysis section focused on the study of the linguistic characteristics of texts is extremely useful in CDA. This allows researchers to better understand how discourse is activated through language (Fairclough, 1992, 1995, 2001). This first level of analysis should be conducted before delving into the second and third level of analysis. However, it should also be conducted during these other two levels of analysis as analysis of discourse as text is highly informative of discursive and social practices. Analysis of discourse as text should be kept alive until the end of the analysis process.

In relation to the preceding recommendation, researchers should not neglect an in-depth understanding of the economic, political, social and cultural context in which they conduct their studies when analyzing discourse as text. The customs, values and beliefs of a given context impact individuals’ ways of being, thinking and acting. These ways of being, thinking and acting manifest themselves through language and, therefore, also in the linguistic characteristics of texts. The context in which texts are produced and consumed should be addressed from the earliest stages of CDA.

We recommend researchers setting a first index of linguistic characteristics to look out for according to their study purposes. These features may respond to grammar, syntax, coherence, vocabulary, punctuation, silences, emotions, flow of ideas, rhythm, associations between events, rhetorical resources, or subjects’ voices in the text (Fairclough, 2001; Wodak & Meyer, 2001). However, we do not recommend sticking with this first index, but to remain open to assess other features not initially considered as relevant. At the end of the preliminary data analysis, researchers will have an index to which they may turn to during the analysis of the codes included in the central and impact and transition sections of each category of data analysis. Analyzing not only what is being said, but how it is being said plays, in fact, a fundamental role in data interpretation.

The analysis of the linguistic characteristics of texts within CDA constitutes, in this sense, a tool to meet criteria of methodological rigor in qualitive research (Morse, 2015). It makes data analysis a more rigorous process given the greater data traceability, since data reporting in sections helps to better understand the data interpretation process, and the greater data transferability, since contextual data analysis has implications for the transposition of research results to similar study contexts.

The analysis of the linguistic characteristics of texts within CDA also constitutes a proof of researchers’ ethical commitment. It places study participants’ identities, experiences, beliefs, and actions at the center of the research process holistically recognizing the scientific value of what they express through them, but also of the way they express them.

Conclusion

This article has aimed to set forth the methodological notes of our experience in conducting CDA. The results of the research we conducted on nurses’ political agency have allowed us to highlight the methodological reasons why discourse as text should be widely acknowledged as a fundamental dimension within CDA. The analysis of linguistic characteristics of texts has been raised as a missed step in nursing research but as a crucial one in discourse analysis aiming to be critical. The analysis of discourse as text in its combination with the analysis of discourse as discursive and social practice has allowed us to be closer to capturing the complexity of what lies behind words. We encourage researchers in nursing conducting CDA studies to adapt the methodological notes highlighted in this article to their own research and to continue producing methodological knowledge in this same direction.