“I Don’t Feel Like I Have Any Control of My Life at All … Everything Overwhelms Me. Everything”: Analyzing Caregiver Uncertainty and Control Through Stance Marking

Abstract

Informal caregivers immersed in the daily care of loved ones at end-of-life stages face such challenges as medical and household issues, worries, doubts, and uncertainties. Using a macro-mezzo-micro approach to discourse, we analyzed parent study interview data involving 46 caregivers facing end-of-life realities. At the mezzo level, we examined caregivers’ expressed perceptions of control. We then more finely analyzed discursive expressions of affective stances pertaining to caregivers’ emotions and feelings, and epistemic stances pertaining to their knowledge and belief states. Theories of uncertainty and control inextricably interweave areas of cognition, affect, and behavior regarding how caregivers perceive their realities and how they engage in or disengage from coping mechanisms in the process. The findings in this three-tiered approach make salient specific discursive patterns gleaned from systematic and fastidious attention to caregivers’ own ways of using language that methodically afford deeper entry into the emotional, physical, and cognitive challenges in their everyday lived experiences.

Introduction

In this article, we present findings from the initial stages of a large ongoing project involving secondary analysis of an extensive set of qualitative parent data on the subject of informal family caregiving for patients at end-of-life stages of their illness. The parent data include more than 400 interviews with 46 caregivers managing the end-of-life care across three distinct trajectories: expected death (i.e., amyotrophic lateral sclerosis [ALS]), unexpected death (i.e., organ failure), and mixed (i.e., lung cancer) (Penrod, Hupcey, Baney, & Loeb, 2011; Penrod, Hupcey, Shipley, Loeb, & Baney, 2012), with interviews conducted roughly once per month for up to 1 year. The interviews were designed to query the caregivers’ impressions and feelings about diagnosis; care management; communication with medical staff; social support; changes in the patient’s physical/emotional condition; changes in overall daily life; and so forth, following a basic protocol of interview questioning. At the same time, the interviewers in all cases both allowed for and encouraged detailed elaboration on the part of each caregiver interviewee. The full basic parent interview corpus consists of 1,309,729 words.

The scope of the initial corpus was narrowed so that we could focus on extended caregiver narratives that address the theme of control, a primary correlate of the concept of uncertainty as examined and analyzed in depth by Hupcey and Penrod (2005), Morse, Hupcey, Penrod, and Mitcham (2002), Penrod and Hupcey (2005), Penrod and colleagues (2011), Penrod and colleagues (2012). In this subset of data, caregivers elaborate dually on those elements and experiences in their daily lives that feel out of control and those elements that help them find or regain control.

Focusing now on a subcorpus of 76,972 words, we conducted a fine-grained discourse-based analysis of the caregivers’ perspectives and viewpoints, with a view to discern with more precision the ways in which these family members (and close friends) characterize their experiences, their loved ones, the disease, themselves, the medical practitioners, and their daily activities. In particular, we analyze their expressions of affective stance, generally conceived as the feelings and emotions associated with the issues and topics, and their expressions of epistemic stance, generally conceived as the knowledge state of the caregivers—essentially what they know and understand, what they do not know and understand, how they investigate issues, what they find out, and where and how they describe still being left in the dark.

As a stance-based, micro-level approach to extended narrative data, we provide insights into the minds and hearts of family caregivers. We analyze the words and transcribed voices of these 46 family members and friends who have now taken on hybridized identities at once as loved one, caregiver, household manager, and oft times untrained participant unexpectedly thrust into the medical community. In concert with Kim and Given (2008), Penrod et al. (2011), and Penrod et al. (2012), we spotlight the words, voices, and perspectives of the informal caregivers in this study.

We outline the various layers of analysis that we employed, from an overall macro-level perspective, to a medium-grained thematic mezzo level, and finally to an initial micro level, which reveals much about the caregivers’ own stances and viewpoints. Our findings dovetail with current research that investigates caregiver burden, both quantitatively (e.g., Sales, 2003; Schene, Tessler, & Gamache, 1994; Schulz & Beach, 1999; Zarit, Reever, & Bach-Peterson, 1980 inter alia) and qualitatively (e.g., Kitt-Lewis, Strauss, & Penrod, 2017; Penrod et al., 2011; Penrod et al., 2012 inter alia), and are expected to contribute to health care research by shedding light on the value of this multilevel discourse analytic approach to health care and nursing.

Research on Informal Caregiving: Statistics and Quality of life (QOL)

Providing care for chronically ill and dying family members has become a socio-familial phenomenon that has been investigated for at least the past four decades in such disciplines as nursing, gerontology, public health, sociology, and psychology (Wolff, Dy, Frick, & Kasper, 2007). As life expectancies lengthen, as costs for hospital and professional care increase, and as medical devices become more and more available for in-home use, so has risen the number of home-based informal caregivers across low- to middle-class U.S. families (Berlinger, Jennings, & Wolf, 2013; Connor & Bermedo, 2014; Hart et al., 2015; Oliver, 2016; Skarf & Schwartz, 2017). As of 2015, the number of informal family caregivers in the United States reached 66 million (National Alliance for Caregiving and American Association of Retired Persons, 2015). This is in sharp contrast with the 1999 statistics reported at 15 million (Schulz & Beach, 1999, p. 2215), or a fourfold increase in just 16 years. By the year 2030, when baby boomers reach their upper 60s and 70s, the issue of informal, in-home care is expected to change even more dramatically than in the decades spanning the 1980s to the present (Knickman & Snell, 2002; Redfoot, Feinberg, & Houser, 2013). From an economic perspective, the estimated value of informal family caregivers’ unpaid support in 2013 ranged between US$470 billion and US$522 billion (Chari, Engberg, Ray, & Mehrotra, 2015; Reinhard, Feinberg, Choula, & Houser, 2015), representing a US$20 billion increase from 2009 (Feinberg, Reinhard, Houser, & Choula, 2011). Current figures are as yet unavailable, but we predict an additional sharp rise in the 5-year span between 2013 and the present.

While informal caregiving may afford major economic savings to families and patients, the practice has taken physical and emotional tolls on the caregiver proper as well as on the broader family unit. Early research on informal care and surrounding care support systems, in general, began in the late 1970s, with the focus initially targeting the elderly and elder care (Cantor, 1975, 1979, 1983; Shanas, 1979; Zarit et al., 1980). Studies on caregiver experiences later expanded to include specific conditions such as Alzheimer’s and dementia (Clyburn, Stones, Hadjistavropoulos, & Tuokko, 2000; Etters, Goodall, & Harrison, 2008; Papastavrou, Kalokerinou, Papacostas, Tsangari, & Sourtzi, 2007), human immunodeficiency virus (HIV), cancer (Kim & Given, 2008), stroke (Tsai, Yip, Tai, & Lou, 2015), chronic illness (Lim & Zebrack, 2004), schizophrenia (Caqueo-Urízar & Gutiérrez-Maldonado, 2006), organ failure (Janssen, Spruit, Wouters, & Schols, 2012), and ALS (Chiò, Gauthier, Calvo, Ghiglione, & Mutani, 2005).

Research assessing QOL and associated caregiver burden, strain, or distress seeks to ascertain specific elements within the caregiving experience that variably aggravate and mitigate perceptions of hardship in the caregiving experience. Causes and degree of caregiver burden or strain have been evaluated from the points of view of gender, relationship to patient, education level, socioeconomic status, age, personality traits and coping styles of the caregiver, objective senses of burden (e.g., daily tasks like bathing, transfer, toileting), and subjective senses of burden (e.g., worry, stigma, guilt) from the overall demands of the job (Adelman, Tmanova, Delgado, Dion, & Lachs, 2014; Cantor, 1983; Faison, Faria, & Frank, 1999; Lökenhoff, Duberstein, Friedman, Costa, & Paul, 2011; Marks, Lambert, & Choi, 2002; Parks & Pilisuk, 1991; Pearlin, Mullan, Semple, & Skaff, 1990; Sales, 2003; Weitzner, Jacobsen, Wagner, Friedland, & Cox, 1999; Weitzner, McMillan, & Jacobsen, 1999; Zarit et al., 1980; Zarit, Todd, & Zarit, 1986 inter alia).

These and numerous other studies of caregiver strain employ a selection of multivariate statistical analytic measures, for example, The Zarit Caregiver Burden Scale and Burden Interview (Zarit et al., 1980), The Caregiver Quality of Life Index (CQOL[C]) scale, the 21-measure of family caregiving burden (Sales, 2003; Schene et al., 1994), Braithwaite’s (1992) measures of caregiving burden, and Caregiver Health Effects Study (CHES, cited in Schulz & Beach, 1999) with a view to more deeply and more specifically identify those areas of caregiving activity that exacerbate stress and those that alleviate it. The instruments are typically designed as self-reporting tools whereby caregivers rank degrees of intensity (using Likert scale-type measures) and/or frequency (e.g., always, often, sometimes, never) of particular feelings associated with caregiving functions (e.g., afraid of what the future holds, comfortable, sadness, frustration, embarrassment). Such research is and has been committed to seeking an overall increased understanding of the various elements that constitute stressors among informal care providers, since as has been amply documented, illness affects not only the patient but also the family members, and especially those who provide care.

Kim and Given (2008, p. 2561) explore conceptual models underlying caregivers’ QOL, with a particular focus on the various stages of the illness (cancer) trajectory. They conclude that caregiver QOL varies over time, from initial diagnosis to bereavement, and underscore the urgency of more in-depth assessment tools to better serve the changing needs of informal caregivers. Kim and Given (2008) examine the stress-coping framework of Lazarus and Folkman (1984) from the point of view of two competing conceptual caregiving models: the wear- and-tear model, where stress increases over time, thereby depleting caregiver’s physical and psychological strength (e.g., Bowman, Rose, & Deimling, 2006) versus the adaptational model, where caregivers gradually adapt and thereby avoid further depletions of strength in the process (e.g., Schulz & Williamson, 1993). While they point out that both models are supported by empirical evidence, the authors conclude that “theoretically and methodologically rigorous research on various aspects of the family’s QOL, including physical, spiritual, and behavioral adjustment to cancer in the family remains sparse” (p. 2563).

More recent qualitative studies on caregiver needs, perceptions, and emotions include Fu, Chen, Li, and Zhu (2018), LaValley (2017), and Lewis (2014), all of which are based on semi-structured, open-ended interviews with guiding questions to channel speakers’ narratives toward the specific issues being queried (hopes, expectations, understanding [Fu et al., 2018]; typical daily activities, changes in caring responsibilities, sources of emotional/social/informational support [LaValley, 2017]; experiences seeking formal end-of-life care/hospice care [Lewis, 2014]). The interviews provide opportunities for caregivers to share their perspectives on their loved ones’ illnesses as well as their own perceptions of hope, despair, certainty, and uncertainty, in addition to elaborating on the types of support, information, and professional interventions that are particularly helpful or unhelpful as they face the ever-changing demands and challenges throughout the course of their role as caregiver. These studies shed important light on the thoughts, feelings, hopes, fears, and coping mechanisms of caregivers facing the potential or imminent deaths of their loved ones. Olson (2016) is a book-length narrative-based treatment of the experiences of 32 spouses caring for their loved ones diagnosed with cancer. Olson provides rich interview data excerpts and personal accounts of these carers, as they share their most intimate feelings and steepest challenges from such basic issues as what it means to be called a “carer” (as opposed to simply “husband” or “wife”) to coping with emotions surrounding the daily routines of caregiving for a spouse, illness, loss, (un)certainty, control and lack thereof, death, and grief.

Concept Analysis, Uncertainty, and End-of-Life Caregiving

Drawing on the methodological and theoretical foundation of concept analysis, Hupcey and Penrod (2005), Morse et al. (2002), and Penrod and Hupcey (2005) establish foundations for an interdisciplinary approach to caregiving in terms of both research and practice. As underscored in Penrod and Hupcey (2005) “the potential of the process [of concept advancement] is to develop a concept toward a more precise scientific definition that permits integration into conceptual frameworks that enhance research or informs or directs practice” (p. 236). Such an approach allows both researchers and practitioners to interweave the elements of phenomena-concepts-practice into the field of nursing. (Hupcey & Penrod, 2005, p. 197). As noted in Penrod (2007), both concept analysis and concept advancement are essential features of concept-based research in nursing, where concept analysis applies to current states of the science and concept advancement builds on and advances knowledge. The recursive interplay between both constructs advance and build theory while also informing best practices. Crucially, concept analysis rests on the quadripartite interweaving of epistemology (relating to the nature of knowledge and the precision of the concept definition), pragmatics (the usefulness of the concept and its “fit” with the discipline), linguistics (the consistent distribution of relevant language use within context), and logic (the integration of conceptual meaning in theoretical formulations) (Penrod, 2007; Penrod & Hupcey, 2005).

Penrod (2007) conducts a concept analysis of uncertainty through a phenomenological caregiver-based study of “living with uncertainty,” leading to both “higher order abstraction” and “greater pragmatic utility” (p. 658). Penrod (2007) complexifies the concept of uncertainty by revealing its essential correlates, that is, confidence and control, such that “strategies for managing uncertainty are more properly based on the critical interplay of the sense of confidence and control rather than identifying or processing discrete probabilities” (p. 662). In other words, the concept analytical study reveals the shift from viewing uncertainty in terms of the classic perspective of “probabilities” to understanding the concept processually as it is lived and experienced on a dayto-day basis by caregivers, as they grapple with the intricately complex ways that past, present, and future events and experiences enmesh into their own perceptions of control and confidence.

Large-scale qualitative applications of the conceptual analytic experience of caregiver uncertainty are reported in Penrod et al. (2011) and Penrod et al. (2012). Both studies were conducted using a database of semi-structured interviews with 46 active adult caregivers (37 females, 9 males) lasting approximately 60 to 90 minutes each, with follow-up interviews designed to be collected once per month for the ensuing 13 months.¹ Of these, 10% of the subjects were care providers to a family member with ALS, 52% to a family member with advanced cancer, and 38% to a family member with organ failure. The time span of active caregiving varied from (pre)diagnosis to death, with the average duration being approximately 2.16 years, and ranging from “a few weeks of intensive activity to years of progressively supportive care” (Penrod et al., 2011, p. 11).

Caregivers for the parent study were recruited from caregiver support groups, palliative care service providers, and a hospital-affiliated senior membership program. Inclusion criteria essentially involved active adult caregivers of adult care recipients with terminal conditions, as those noted above. There were no explicit exclusion criteria. The majority of the caregivers were spouses (64%). All of the male caregivers were spouses; female caregivers included spouses and other familial relationships. The caregivers ranged in age from 36 to 84 years (mean age = 62). The care recipients ranged in age from 36 to 93 (mean age = 71). Data were collected in face-to-face interviews at the care recipients’ homes and also by telephone. All interviews were transcribed verbatim and then verified by the researchers to ensure accuracy of the data sets and transcriptions.

In light of the unstructured nature of the interviews and the liberal time designation allotted for each participant’s contributions, the interactional process provided space for caregivers to recount their personal experiences, often taking the form of extended narratives, in response to selected open-ended questions or prompts (e.g., “How have things changed since your [family member] was diagnosed?,” “Was there ever a time when you felt like things were ‘out of control’?”). The format and methodology behind the data collection process enabled the researchers to analyze the responses using Grounded Theory (Glaser, Strauss, & Strutzel, 1968) and uncover poignantly personal and analytically powerful themes in the three basic end-of-life caregiving trajectories: expected death trajectory (ALS), mixed death trajectory (cancer), and unexpected death (organ/heart failure).

The unified theme in both studies is that of “seeking normal,” in which caregivers recount their experiences throughout the recent life history of their family member’s (pre)diagnosis, treatment, improvement/decline, and in most cases from the data sets, death. Penrod et al. (2011) and Penrod et al. (2012) uncovered thematic patterns in the 46 caregivers’ discourse that converged into the following emergent themes: Sensing a Disruption, Challenging Normal, Building a New Normal, and Reinventing Normal. In their responses, caregivers expressed their perceptions, fears, hopes, and challenges throughout the process. The concept of uncertainty as a lived experience in the day-to-day cycle of family caregiving emerged in relation to the associated correlates of confidence and control.

Penrod et al.’s (2011; Penrod et al., 2012) discussion of uncertainty and control resonates with Mishel’s (1988, 1990) theory of uncertainty in illness on the one hand and Bandura’s (1997) theory of self-efficacy as control on the other. Mishel’s uncertainty in illness theory centers on the cognitive (in)ability of patients (and caregivers) to comprehend illness-related events (e.g., diagnosis, treatment options, prognosis) coupled with how they structure meaning for those events. Uncertainty manifests in the areas of ambiguity about the state of illness/wellness, insufficient or abstruse information, and unpredictability of treatment, prognosis, complications. Mishel’s (1988) theory posits two cognitive response patterns of appraisal that emerge as reconciliatory or adaptive means vis à vis uncertainty: inference, by evaluating uncertainty based on lived experience, belief systems, and education; and illusion, by evaluating potential for positive outcomes. In each case, appraisal leads to various styles of coping mechanisms that individuals appeal to, alone and in concert with other sources of support, as they strive to develop new patterns of behavior, reformed understandings, and a renewed sense of control, in the ongoing flux of time and illness, day-by-day, month-by-month, year-by-year. In fact, Mishel’s (1990, p. 258) expansion of his original theory addresses this crucial intersection of time and its evolutionary influence on the individual patient and carer; on fluctuating states of illness, remission, and wellness; on the contexts and environments of the individuals’ daily lives; and on the ever-changing and evolving mechanisms of uncertainty appraisal. Within this expanded theory, the focus on cognitive, emotional, and behavioral adjustment shifts from an “equilibrium ideal” to states of increased complexity, yielding more possibilities and “new patterns of contingencies,” hence control (Mishel, 1990, p. 261).

Bandura (1997) puts forward a cognitive theory of control that is grounded in the notion of self-efficacy. Essentially, if individuals believe they have the ability or power to bring about particular outcomes, they will engage in activities geared toward achieving those results. Conversely, if they believe they have no such power or no such ability, they will not. The theory intertwines the domains of the cognitive, the affective, and the behavioral, all of which underlie both imagined and implemented coping mechanisms as well as human agency more generally.

In a similar vein, in their work with the elderly, Piper and Langer (1986) put forward their notion of mindfulness in relation to control or more specifically in relation to the agentive process of achieving a particular desired result. Cognitively, mindfulness relates to the “process of differentiating, drawing distinctions, creating categories, or making meaning.” Its negative correlate, mindlessness, is “a state of reduced cognitive activity that results from a reliance on rigid structure that may arise after several repetitions of an experience” (p. 72). Mindfulness, as a dynamic process, leads to control. Mindlessness, as a state, involves automatic modes of apprehending and evaluating control on the basis of outcomes viewed dualistically as simply success or failure. Failure, from a static observer’s perspective, is the negative outcome of a single attempt, repeated and reified as an endpoint result. In contrast, a mindful approach centers on the actor and the actor’s own perspective, where nonaccomplishment of a desired task is simply an inevitable step within the larger process of achieving goals, one small one at a time.

Such studies contribute to nursing praxis in that they provide deeper insights and keener understandings of caregivers’ needs throughout the changing phases in “one of life’s most difficult journeys—ushering the death of a loved one” (Penrod et al., 2012).

Caregivers’ Voices and Stances—Discourse as the Broad Lens for Secondary Analysis

In this article, we present results of a study that takes a macro-mezzo-micro discourse analytic perspective on the parent data, that is, Exploring the Formal/Informal Caregiver Interface Across 3 Death Trajectories (National Institutes of Health (NIH)/NINR R01NR010127, Penrod et al., 2011). All Institutional Review Board (IRB) clearances and permissions were obtained for all three authors. The entire corpus of parent data, having been cleared of any identifying references to people and places, was reviewed on multiple occasions by all three authors for use as secondary data in this phase of the large-scale discourse-based research.

We aim to elucidate in depth Morse et al’s (2002) four-component framework for concept analysis (epistemology, linguistics, pragmatics, and logic) focusing primarily on the linguistic and pragmatic elements and to concomitantly continue advancing the inextricable interplay of concept, theory, and nursing praxis. The findings will also be discussed from the perspectives of Bandura (1997) and Piper and Langer (1986) on control, self-efficacy, and mindfulness and Mishel (1988, 1990) on uncertainty.

The corpus is rich with personal narratives by family members (and close friends) charged with the care, comfort, and in-home treatment of their loved ones. Given that the narratives are essentially elicited to chronicle the various and changing stages of the three end-of-life trajectories (expected death, unexpected death, and mixed) through the caregivers’ viewpoints and feelings about the patient and themselves, they are imbued with feelings, hopes, thoughts, despair, and doubts. As a secondary data corpus, we analyze these narratives through macro-mezzo- and micro-level lenses of discourse, uncovering and examining the patterns created by the caregivers’ choices of words, response types, repetitions, reformulations, questions, statements, and so forth.

Discourse is defined as the

social and cognitive process of putting the world into words, of transforming our perceptions, experiences, emotions, understandings, and desires into a common medium for expression and communication, through language and other semiotic resources.

(Strauss & Feiz, 2014, p. 1)

From this perspective, an analysis of discourse has the potential to reveal the multiple complex modes in which speakers perceive themselves, their interactants, and their environment through their choices of nouns, verbs, pronouns, complex or simple sentence structures, repetitions, and other linguistic markers. The analysis put forward here rests on the in-depth examination of expressions that mark the speakers’ stances or feelings, positions, opinions, and attitudes vis à vis the issues under discussion.

Discourse and Stance

We begin the discourse-based analysis with a focus on the caregivers’ stances as expressed in their narrative responses to the interviewers’ questions. Stance has been discussed in linguistics literature as an indicator of point of view (Chafe, 1994) and evaluation (Conrad & Biber, 2000). More broadly, Positioning Theory (Harré, 1999) in psychology centers on the ways in which language use points to dynamically emergent constructions of selfhood and otherhood within context and social interaction. Stance, then, essentially refers to the manifold array of linguistically captured evaluative perceptions that emerge in discourse (Biber, Johansson, Leech, Conrad, & Finegan, 1999; Chafe & Nichols, 1986; Du Bois, 2007; Goodwin, 2007; Johnstone, 2009; Ochs, 1996; Ruusuvuori, 2013 inter alia). Building on this foundational work on stance, Strauss and Feiz, 2014 define stance as follows:

Stance is the speaker’s or writer’s feeling, attitude, perspective, or position as enacted in discourse. Stancetaking is an inevitable consequence of participating in and producing discourse, of putting the world into words. Stance emerges in a speaker’s or writer’s choice of one linguistic form over another, the coloring of utterances with prosodic contours or punctuation, the sequential ordering of utterances; it emerges in gestures, silences, hesitations, hedges, and in overlapping stretches of talk. In all of these instances of discourse (and others), a speaker’s or writer’s stance is enacted and created, it is negotiated and re-negotiated.

(p. 4)

Categories of stance-marking in discourse include but are not limited to the following: affective stance (Du Bois, 2007; Du Bois & Kärkkänen, 2012; Ochs, 1996; Ruusuvuori, 2013), epistemic stance (Biber et al., 1999; Du Bois, 2007; Ochs, 1996; Ruusuvuori, 2013), cooperative stance (Goodwin, 2007), moral stance (Goodwin, 2007), and authoritative stance (Johnstone, 2009). For the purpose of the present analysis, we focus the discussion on affective and epistemic stance.

Affective stance relates to “a mood, attitude, feeling, and disposition, as well as degrees of emotional intensity vis-à-vis some focus of concern” (Ochs, 1996, p. 410). As such, it is associated with a speaker’s or writer’s emotions and feelings. Linguistic markers of affective stance include nouns, verbs, and adjectives that relate to feelings (e.g., fear, happy, ecstatic, relieved, overwhelm) in addition to intensifying adverbs (e.g., very, really, horribly).

Epistemic stance, in turn, differs from affective stance in that it rests not within the domain of feeling but rather in the domain of thinking, logic, cognitive processing, and understanding. In parallel with affective stance being related to degree of emotional intensity, epistemic stance is also associated with degree of certainty or uncertainty vis-à-vis an issue at hand. Markers of epistemic stance include verbs of cognition (e.g., know, understand, believe, guess, wonder, surmise, doubt), epistemic modals (e.g., must/should—“Marta must/should be home by now;” might/may/could—“It might/may/could rain today”), nouns, adjectives, and adverbs that relate to degrees of (un)certainty (e.g., maybe, certainly, of course, (un)sure, (un)certain, doubtfully, undoubtedly, probably).

The focus of this study is on the instantiations of epistemic and affective stance by the caregivers in their extended narrative responses to the interviewers.

Research Questions

Using the more fine-grained lens of discourse analysis, with a focus on affective and epistemic stance, our study sets out to investigate the following research questions:

• Research Question 1: Specifically, how can a discourse analytic perspective elucidate the concept of uncertainty, with an initial focus on control?

• Research Question 2: What are the sources of losing control as reported by interviewees?

• Research Question 3: What are the sources of gaining control/relief as reported by the caregivers?

• Research Question 4: How do the concepts of affective and epistemic stance help to disambiguate the feelings/perceptions of caregivers?

Method

Macro-Level Observation: Theme of Control

To methodically probe the large-scale parent data set, the authors followed a set of systematic procedural steps upon IRB approval. The authors reviewed the full set of transcripts to ensure that any reference to participants and to their family members was fully and unequivocally anonymized and reference to other potentially identifiable persons and places were removed or altered to ensure non-identifiability.

The transcripts were then compiled into the first-level corpus consisting of the full sets of interviews between interviewer and caregiver. The total number of words in the overall corpus is 1,309,729, as indicated in Table 1. Word counts in corpus-based discourse studies are important in that they reveal the “size” of the database, measured in number of words. Quantitatively, the total number of words in each corpus reveals relative size distinctions in addition to providing potential ground for the analysis of patterns and frequencies of specific word forms and collocations-the natural, patterned co-occurrence of a particular word with one or more other words in a corpus. With regard to relative corpus size, Table 1 provides the total number of words in the full data set, and Table 2, the total number of words in the more thematically focused corpus.

Table 1.

Total Words in Full Corpus According to Trajectory.

Trajectory	No. of Words
Expected death (ED)	363,751
Unexpected death (UD)	480,102
Mixed (M)	465,876
Total	1,309,729

Table 2.

Subcorpus—Theme of Control.

Trajectory	No. of Words
Expected death (ED)	22,283
Unexpected death (UD)	31,187
Mixed (M)	23,502
Total	76,972

We initiated a basic macro-level thematic review of the full transcript sets by first randomly selecting one full transcript set from each of the trajectories, including all transcripts from the initial meeting to the final meeting. We then reviewed the transcript subsets for overall content with a view to gain a preliminary sense of the types of linguistic patterns in the data sets of the basic full corpus. Through this process, we identified the single lexical item control as a salient linguistic marker that would be worthy of deeper investigation, especially in light of the foci of the parent studies, that is, “seeking normal” and “living with uncertainty” (Penrod et al, 2011; Penrod et al., 2012). The term emerged in some of the interviewers’ lines of questioning and in all the caregivers’ responses.

We were thus able to trim the size of our initial working database from the full basic interview corpus to now include only those full episodes centering on questions and responses that contained the word control. The subset of data encompassed all 46 respondents’ interviews. In each case, the episodes involved narratives of caregivers’ senses of losing control, feeling out of control, gaining control, and so forth, which were often, but not always, prompted by such questions as in Excerpt (1), from the baseline interview protocol:

• 1Excerpt from baseline interview protocol
  Another area that seems to be of concern to caregivers is feeling in control of situations. Was there ever a time when you felt like things were ‘out of control?’

  • Can you tell me about that experience?
  • What was going on at the time?
  • What were you thinking during this time?

The narrowing, thematically driven process of data management allowed us to then create a thematic subcorpus, in which the theme of control is central. The size (in number of words) of the subcorpus and each of the three trajectories are indicated in Table 2.

Mezzo-Level Analyses: Sources and Issues Surrounding Lost and Regained Control

The thematically central subcorpus of control allowed for a deeper and more systematic investigation into the concepts along the spectrum of not having control of an issue (having less control, feeling out of control, losing control) at one end of the spectrum and having control (finding/regaining control, experiencing relief) at the other.

It is likely that the polarized perspectives of not having control versus having control were in part influenced by the line of questioning of the interviewers; however, it must be noted that not all interviewers actually posed these questions explicitly. In some cases, the topic simply emerged unprompted in the caregivers’ narratives.

From the 76,972-word subcorpus of interview/narrative data, we further categorized the caregiver responses into patterns identifying specific sources and issues for underlying the more negative categories of lack of control and the more positive categories of having control, respectively.

For the categories of lack of (less, losing) control, we identified four sources in the caregiver data, separated according to where the focus is placed and the areas in daily life where the caregiver perceived a lack/loss of control. We further found four major vectors that were associated with loss/lack of control: the patient (physical condition, psychological state and mood, medication management), the caregiver him or her self (psychological outlook, physical ability, and appearance), medical issues (medical staff, treatment, and support), and issues that fall outside of the prior three (such as household responsibilities, insurance, family support). These are summarized broadly in Table 3.

Table 3.

Factors Associated With Not Having Control.

Source	Specific Issues
Patient focused	Physical conditions, psychological state/mood/outlook, medication management
Self-focused	Psychological outlook, physical ability/fatigue/appearance
Medical	Medical staff/treatment issues, medical support
Outside (beyond patient, medical care, or self)	Household roles/responsibilities, insurance, social/family support, other responsibilities

For the categories of having and/or gaining control, that is, sources of regaining control or source of relief, four primary vectors were identified: sources outside the caregiver—particularly other individuals, including the patient (mood, quality time spent with patient, social support, medical support); institutional sources of relief (government programs, financial support); source of relief or control from caregivers’ choices (physical activity, exercise, temporarily leaving the home environment, deciding to relocate residence to accommodate patient’s needs); sources of relief/control from caregivers’ experience or knowledge, psychological outlook, or faith (outlook on life, experience with and/or knowledge of the disease, religious faith). See Table 4.

Table 4.

Factors Associated With Having/Regaining Control, Sources of Relief.

Source	Specific Issues
Source of relief/control from outside the caregiver	Patient mood, quality time with patient; social support/family; medical staff/medical support
Institutional source of relief	Government programs, financial support
Source of relief/control from caregiver choices	Physical activity, exercise; temporarily leaving the home environment; relocating residence to better accommodate patient
Source of relief/control from caregiver experience, knowledge, psychological outlook, faith	outlook on life, experience with and/or knowledge of the disease or condition, religious faith, prayer, trust in God

The general results of this thematic, mezzo-level analysis point broadly to concepts of uncertainty in illness (Mishel, 1988, 1990; Penrod, 2007; Penrod et al., 2012), degrees of mindful and mindless coping strategies (Piper & Langer, 1986), and strategies that point to varying degrees of the caregivers’ perceived self-efficacy in terms of agency and control (Bandura, 1997; Penrod, 2007; Penrod et al., 2012).

Micro-Level Analysis: Epistemic and Affective Stance Marking

At the micro level of analysis for the subcorpus of control, we focused on affective and epistemic stance-marking, examining in detail the events, situations, interactions, and perceptions that caregivers characterized as positive and those characterized as negative.

Positive affective stance-marking includes such evaluative terms and expressions as good, helpful, kind, happiness, relief, joy, and so forth. Affective stance is also indicated indirectly or implicitly through expressions of high agency, where speakers or writers indicate their ability to change and/or exert control over an issue at hand, thus expressing positive affective stance, for example, “I just walk away from it, and he’ll cool off” (UD). Agency is defined skeletally by Ahearn (2001, p. 112) as “the socioculturally mediated capacity to act,” and by Duranti (2004, p. 453) as “the property of [an individual] … that [has] some degree of control over their own behavior, and … whose actions in the world affect other entities … ”

Affective stance is also evident in expressions that point to emotive reactions vis-à-vis a particular situation without actually using an emotive term. Caregivers’ expressions of affective stance and agency correlate well with Bandura’s (1997) concept of self-efficacy, in that when caregivers spontaneously recount narratives in which they exercise agentive behaviors leading to positive affective stance-marking, they manifest an underlying belief that they have some degree of power over their immediate situation in relation to their coping strategies.

Extract set (2) illustrates the patterns in caregiver data that express the caregivers’ positive affective stance:

• 2
  Illustrative extracts from caregiver narratives: affective positive stance marking

  1. We take every day and roll along with it. If we’re here the next day, we’ll roll along with that one and we’ll change as we need to go (ED)
     [positive orientation toward coping with situation, high agency]
  2. As long as I feel I have control of a situation, I feel pretty cocky. (M)
     [having control makes caregiver feel extremely confident, high agency]
  3. So I think our spending time together [as a couple] is actually what alleviates my concerns and my fears and helps to pull me up a little bit. I think it does him as well. (UD)
     [activity with partner makes caregiver feel better, improved psychological outlook and overall sense of comfort]
  4. I do look forward to seeing you [you = interviewer] because it allows me to get my thoughts out to try to express them, when you’re just thinking about them sometimes they’re just a jumble in your mind. You go over them here and there, but I think trying to express them in words vocally, allows my thought process to work better. (UD)
     [interaction with interviewer makes caregiver feel better, improved psychological outlook and overall sense of comfort]

Conversely, negative affective stance-marking includes such expressions as difficult, disappoint, overwhelm, upset, worry, scary, nasty, and so forth. It is also evident in expressions of low or no agency in the sense that the individual comments on an inability to change and/or exert control over an issue, as in the following statement “ … having the kitchen counter looking neat and yeah, I feel like that stuff gets away from me … ” (ED). Negative affective stance-marking from the caregiver’s narratives is exemplified in (3):

• 3
  Illustrative extracts from caregiver narratives: affective negative stance marking

  1. Yea, and I think I don’t feel like I have any control of my life at all. I don’t have control. Everything overwhelms me. Everything. From the house, everything is up to me. All this, decisions, nothing, I can’t even help (M)
     [negative feeling derives from caregiver’s sense of having no control, low/no agency, feeling helpless]
  2. I think a lot of it is just underlying pressure again about the future … I’ll lay in bed tonight and worry about, you know the big things that you depend on often times your spouse to take care of. (UD)
     [feelings of worry and pressure deriving from an unpredictable future]
  3. No, the problem is, not being able to boom, boom, boom, boom, boom. Get things done and over with. Everything’s like wait 3 weeks, wait 4 weeks, wait 3 months. This is the way it’s been. It really gets to me. (UD)
     [feeling of frustration deriving from caregiver’s inability to take care of things swiftly, low agency]
  4. d.… I’m a worrier and I let all these little things, the what if’s, go through my mind all the time. I worry about that. (ED)
     [negative feeling deriving from caregiver’s proclivity to naturally and constantly worry]

In the case of epistemic stance-marking, positive epistemic stance is evoked through verbs of cognition—often in the affirmative, for example, know, understand, figure out, while negative epistemic stance is evoked through similar verbs expressed with negative morphosyntax, for example, don’t understand, can’t figure out, and so forth. In addition, positive/negative epistemic stance-marking includes adjectives that relate to knowledge systems and general understanding, such as certain/uncertain, sure/unsure, and so forth. While at times the two categories of affective and epistemic stance appear to overlap, epistemic stance-marking emphasizes the cognitive element involved in the response, and not solely (if existent at all) the affective side. Furthermore, epistemic stance-marking is evident in the speaker’s expression of degrees of certainty, often with epistemic modals will, may, might, for example, “I might be able to figure it out” (possibility) versus “I will find out more about his condition tomorrow when I press the doctor.” Sometimes, absolute certainty is expressed without concomitant modal verbs or adverbs, for example, “Be careful. This must be a trap” (modal must = high probability) versus “Be careful, this is a trap” (no modal = absolute high certainty.) (Strauss, Feiz, & Xiang, 2018).

The extracts in (4) and (5) illustrate positive and negative epistemic stance-marking, respectively:

• 4
  Illustrative extracts from caregiver narratives: epistemic positive stance marking

  1. I don’t think I’m riding on the edge as to where I’m going to explode. I’ll figure something out (ED)
     [caregiver expresses confidence in cognitive ability to address the issue]
  2. So I haven’t lost the confidence that I can do the job. I’m just a damn stubborn person. [Laughter] I know what I can do and I know what I can’t do. This I can do! (UD)
     [caregiver expresses confidence in cognitive ability to address the issue]
  3. I actually think probably in some cases with his medical coverage and with the studies and things like that he’s probably getting excellent care in comparison to maybe people who don’t have health care. So I would say he’s probably getting as good of care as he could possibly get I guess. So in that respect I’m comfortable knowing that. I know there’s not much more that I could do or anybody else you know to help his situation. (M)
     [caregiver expresses confidence in cognitive ability to address the issue]
• 5
  Illustrative extracts from caregiver narratives: epistemic negative stance marking

  1. I can’t get him to the hospital any quicker you know. When you sit at Cedars [pseudonym] Medical Center and you’re waiting all day for a doctor you know. You know they’re doing what they need to do; you know they’re reviewing tests; you know they’re doing what they need to do but you’re still sitting there for you know 6, 8, 10 hours or whatever waiting to find out what will the end result be—what happened or you know where do they have to go from here. (UD)
     [caregiver’s lack of knowledge regarding patient’s condition] b … the only time I get down is with this obsessive thinking you know when I feel a little unsure of myself or I don’t have answers or something like that then. I’m a little insecure I guess you would say, that’s part of the situation. (M)
     [caregiver’s lack of knowledge regarding patient’s condition]
  2. I’m a little uncertain if I’m going to be able to be a good caretaker in the sense that the wheelchair, if I can put it in and out of the car. It’s going to be heavier than what I can handle. I’m not going to be able to lift a heavy, if it’s more than 20 pounds, there’s no way I can lift it. (ED)
     [caregiver’s lack of certainty in her own ability to accomplish task]

These mezzo- and micro-level instances of discourse that reflect narrators’ epistemic viewpoints, stances, and perspectives will shed light on Mishel’s (1988, 1990) and other scholars’ work on the various instantiations of uncertainty in relation to chronic illness and how illness affects the unfolding of daily life over time, in that it will allow researchers to more precisely identify broader instances of inference and illusion and how these concepts change longitudinally.

Findings: Epistemic Commentary and Affective Commentary

By establishing stance categories as initial units of analysis, we have gained more fine-grained insights into caregivers’ feelings and attitudes with respect to themselves, their loved ones, the disease, the medical care, and their future lives. Tables 5 through 8 present the quantified results of caregivers’ expressions of affective and epistemic stance. Each table reflects the seven categories of focus that emerged in relation to the caregivers’ expressions of both affective and epistemic stances: patient, self, situation of illness in general, dyadic relationship (e.g., spouse-spouse, parent-child), medical issues, social factors (family, friends, groups), and religion.

Table 5.

All Trajectories: Negative and Positive Commentary.

	Epistemic Commentary		Affective Commentary
Focus	Positive	Negative	Positive	Negative	Totals
Patient			15	40	55
Self	85	14	41	92	232
Situation			2	34	36
Dyadic relationship	3		3	4	10
Medical	6	1	24	29	60
Social: family, friends, groups			25	4	29
Religion			12		12
Total	94	15	122	203	434

Note. Epistemic positive: Epistemic negative 94:15 = 6:1. Affective negative: Affective positive: 203:122 = 1.6:1.

Table 8.

Expected Death Trajectory: Negative and Positive Commentary.

	Epistemic Commentary		Affective Commentary
Focus	Positive	Negative	Positive	Negative	Total
Patient			3	10	13
Self	28	4	25	31	88
Situation			2	14	16
Dyadic relationship	2				2
Medical	1		7	2	10
Social: family, friends, groups			11	4	15
Religion			4		4
Total	31	4	52	61	148

Note. Epistemic positive: Epistemic negative 31:4 = 2.9:1. Affective negative: Affective positive: 61:52 = 1.2:2.

Our findings reveal that by virtue of the affective/epistemic stance distinction, overwhelmingly, epistemic positive commentary (e.g., “I’ll figure something out”) is far greater than epistemic negative commentary (e.g., “ … I’m a little uncertain if I’m going to be able to be a good caretaker … “), across all three trajectories. Furthermore, affective negative commentary (e.g., “Everything overwhelms me. Everything”) outnumbers affective positive commentary (e.g., “our spending time together is actually what alleviates my concerns and fears”), though the ratio skew is consistently smaller than that of the epistemic category. Recall that the excerpts in (2) to (5) illustrate instances of positive and negative commentary: affective positive (2), affective negative (3), epistemic positive (4), and epistemic negative (5).

Table 5 presents the results from the combined three trajectories: Epistemic positive commentary is 6 times greater overall than epistemic negative commentary. Affective negative commentary is 1.6 times greater than affective positive commentary. In the combined trajectories (Table 5) and in the individual trajectories for mixed death (Table 6), unexpected death (Table 7), and expected death (Table 8), the greatest number of epistemic positive references is in relation to the caregivers themselves, noted in the category “self.” The greatest number of affective negative references is also in the category of “self.” The skews concerning caregivers’ perceptions of “self” are highlighted in bold in each table.

Table 6.

Mixed Death Trajectory: Negative and Positive Commentary.

	Epistemic Commentary		Affective Commentary
Focus	Positive	Negative	Positive	Negative	Total
Patient			10	8	18
Self	37	8	8	28	81
Situation				14	14
Dyadic relationship	1		1	1	3
Medical	4		10	16	30
Social: family, friends, groups			9		9
Religion			6		6
Total	42	8	44	67	161

Note. Epistemic positive: Epistemic negative 42:8 = 5.25:1. Affective negative: Affective positive: 67:44 = 1.5:1.

Table 7.

Unexpected Death Trajectory: Negative and Positive Commentary.

	Epistemic Commentary		Affective commentary
Focus	Positive	Negative	Positive	Negative	Total
Patient			2	22	24
Self	20	2	8	33	63
Situation				6	6
Dyadic relationship			2	3	5
Medical	1	1	7	11	20
Social: family, friends, groups			5		5
Religion			2		2
Total	21	3	26	75	125

Note. Epistemic positive: Epistemic negative 21:3 = 7:1. Affective negative: Affective positive: 75:26 = 2.9:1.

That caregivers’ responses exhibit a remarkable skew in the area of epistemic positive stance-marking (with a focus on “self”) sheds much light on how they perceive themselves as persistent and capable problem solvers. We find similar patterned responses in the area of affective positive stance-marking, as caregivers make affective connections between particular situations in their lives and how those situations help them feel better, uplifted, and confident. Participants’ discourse from both categories of positive epistemic and positive affective stance-marking reflects features of perceived self-efficacy and of a perceived and actual sense of agency, in that they recount episodes of having taken positive steps toward effecting change and/or comfortably accepting a situation knowing that they have made the best possible choices and that their loved one is in the best possible care.

Conversely, the skew in the area of affective negative stance-marking (with a focus on “self”) points to a preponderance of feelings of helplessness, pressure about the future, and worry. Epistemic negative stance-marking evokes uncertainty and self-doubt in the sense of not having answers and not being able to predict an outcome—whether it is related to treatment, prognosis, and even to the caregivers’ own physical ability to accomplish a care-related task.

Throughout the data, we find consistent reference to the caregivers’ cognitive perceptions of their ability to understand and carry out a task, intermeshed with reference to their affective reactions in terms of how these perceptions make them feel. Often caregivers relate these cognitive and affective perceptions to actual or imagined outcomes or behavior.

In this sense, the determination of elements in the narratives that evoke the caregivers’ affective and epistemic stances afford researchers fresh perspectives into the original concept of uncertainty. That is, in line with Penrod (2007) and Mishel (1988, 1990), rather than viewing uncertainty as pertaining squarely to the category of negative epistemic stance (or the traditional notion of “probabilities”), this analysis complexifies the concept as it is lived and experienced by the caregivers on a day-to-day basis, linking epistemic/cognitive experiences with the affective and also the behavioral.

More Fine-Grained Micro-Level Analysis: Individual Word Tokens and Frequencies

The analysis now proceeds further to more fine-grained micro-level uses of language affording a glimpse into the ways in which the caregivers in this subcorpus used the adverbial intensifier very to underscore an evaluative assessment of elements in their daily lives of caregiving. Recall that Ochs’s (1996) definition of affective stance includes the “degree of emotional intensity” expressed by a speaker or writer with regard to an issue being evaluated, judged, or assessed. Within the subcorpus of the control-themed interview segments, we noted 82 total tokens² of the intensifier very. Of these, eight were categorized as “peripheral” in that they do not serve to intensify a stance that directly relates to the caregivers’ day-to-day issues of managing their loved one’s care. In other words, they occur in secondary or peripheral narratives not central to the patient and/or the caregiver proper.

Of the remaining 74 tokens, 31 (42%) emphasize a positive evaluative term such as good (seven tokens), helpful (four tokens), well (three tokens), independent (two tokens), proud, blessed, faithful, great, and accommodating. In contrast, 43 tokens (58%) co-occur with a negative evaluative term such as upset (five tokens), not very well (four tokens), not very much (three tokens), difficult (two tokens) lonely, rough, scary, tenuous, disappointed, distressing, and hurtful.

These results appear in Table 9.

Table 9.

Instances of Intensification Using Very in the Control Subcorpus.

Peripheral, Positive, Negative Uses of Very	Focus	Of tokens	Examples
Peripheral (secondary narratives)		(8)	Very rarely do I listen to it (TV news) She was with him at the very end.
Positive	Patient	8	… very healthy for someone with lung cancer She was very proud of being a cancer survivor.
	Self	8	We get along very well, I am very blessed … At home, I can handle it very well. I just walk away …
	Medical	9	We always have very good access to the care manager … her job. She is very good at it.
	Social	6	He (the priest) is very faithful … The people (at church) are just great, very great.
Subtotal positive			31/74 = 42%
Negative	Patient	19	It was very difficult to understand him. He wasn’t thinking very good thoughts. And that scares me.
	Self	22	I was very upset about that. I am not very good at asking for help. I’ll tell you what it is: lonely, very lonely
	Medical	1	The doctor was actually amazed and shocked and was very disappointed.
	Situation	1	… you have days where everything is just piling up, you know, like we picked a very chaotic day that day.
Subtotal negative			43/74 = 58%
Subtotal positive and negative		74
Total tokens of very		82

Other linguistic markers in the data that evoke affective stance include intensifying adverbs such as really (“plus everything is really bland”), totally (“ … you are totally helpless”), and pretty (“I think things are going pretty well). We have examined these and other discursive patterns for next stages in the analysis, patterns such as hyperbole and exaggerated imagery (“ … like a tornado went through my house”), self-directed rhetorical questions (“I thought, ‘How are we going to do this?’), and repetitions (“Everything overwhelms me. Everything”) (here, repetition plus hyperbole), for additional perspectives to the caregivers’ cognitive/affective orientations toward their challenges and realities.

Discussion, Limitations, Future Applications, Implications for Health Care

This study set out to answer the four research questions posed in the “Discourse and Stance” section and is intended as a foundation for and supplement to current ongoing work that targets the nursing discipline specifically as well as the fields of linguistics, applied linguistics, and linguistic anthropology more generally.

In line with the more complex view of uncertainty as a concept that expansively transcends the classically more simple construct of “probabilities,” pointing to the singular domain of the unknown, we have demonstrated through macro-, mezzo-, and micro-level discursive patterns that the concept of uncertainty in these end-of-life caregiver narratives encompasses feelings of control and lack of control that resonate with psychological theories of self-efficacy and agency (Bandura, 1997) in addition to the linguistic anthropological concept of agency (Ahearn, 1999, 2001; Duranti, 2004; Strauss & Feiz, 2014; Strauss et al., 2018). That said, we note a strong correlation in the caregivers’ discourse between control as used in the parent study and agency, in the contexts of both epistemic and affective stance-marking.

The teasing apart of these two powerful constructs of stance (i.e., epistemic stance and affective stance) allows researchers to more precisely understand and more systematically complexify the construct of uncertainty (and its correlates of both control and confidence) and to gain more fine-grained insights into the thoughts, beliefs, fears, hopes, and understandings of family caregivers. The constructs reveal aspects of what the participants do or do not think, believe, and understand about themselves, their loved ones, and the illness itself, on the one hand, and how they feel about these issues on the other, in addition to how such emotions and beliefs play out in behavior—both real and imagined. The fact that positive epistemic stance-marking discernibly outweighs stance-marking in the other three categories points to the finding that the caregivers in this data set did possess positive beliefs about their abilities to cope and manage particular elements within their caregiving experiences at particular stages of the illness. Perceptions of control and self-efficacy, as expressed in the positive epistemic and positive affective categories of stance-marking, are relatively higher than perceptions of worry, uncertainty, doubt, fear, and so forth. Expressions of negative affective stance-marking, while higher than positive affective stance-marking, reveal specific types of feelings that the caregivers experience in their daily lives and those that concern them about the future.

The perspectives put forward here also add granularity to the existing grounded-theoretical/phenomenological qualitative work as well as to the more voluminous earlier multivariate statistical analytic research that queries caregiver burden and stress using Likert-type scales to identify the feelings and perceptions that caregivers experience and the general frequencies of occurrence that these feelings emerge.

The limitations of this study include the fact that our findings are based on a thematic subset of the original 1,309,729-word parent corpus. The narrowed frame of inquiry allowed us to design an analytic framework together with basic analytic units and theoretical support to inform the discursive findings. A next step in this project involves the extension and application of the framework to the full original data set. Also, because the parent study was designed to illuminate longitudinal changes over the intended 13-month data collection period, the results lend themselves to a longitudinal granular perspective of change over time, which exceeded the scope of the current project. In addition, a closer examination of the caregivers’ narratives across the three death trajectories will also yield more insights into these issues. Furthermore, the parent data set potentially lends itself to an extended secondary study that applies a fine-grained lens to Mishel’s (1988, 1990) theory of uncertainty in illness, with a specific focus on inference-based and illusion-based response patterns of appraisal—discursive elements that could also be identified and tracked over time for evidence and quality of change.

Future researchers and practitioners in the arenas of hospice, palliative care, and caregiver support can begin to attend to specific ways in which both caregivers and recipients of care express their understandings, beliefs, and feelings. Interventions can be designed to include an array of patterned target expressions that practitioners can listen for and learn to address in such a way that caregivers’ and patients’ perceptions of self-efficacy are recognized and responded to, with a view to continue building on and bolstering extant perceptions of positive epistemic and affective stance.

Author Biographies

Susan Strauss is assoicate professor of Applied Linguistics/Asian Studies at Penn State University. Her research interests include discourse analysis; cognitive linguistics; second language writing; intercultural communication; media discourse; and the interface between discourse, grammar, stance, culture, and communication.

Erin Ann Kitt-Lewis is assistant research professor at Penn State University with research foci in end-of-life, caregiving, and community-based research.

Michael Amory is a PhD candidate in Applied Linguistics at Penn State Universtiy. His primary research interests include: the theory and practice of second language teacher education and utilizing the framework of Conversation Analysis to analyze classroom interactions and institutional discourse.