2.1. Simple and narrative testimony

Testimonial compression occurs when the audience of a testimonial exchange enforces the format of the testifier's engagement, specifically enforcing the format of simple testimony. As defined by Rachel Fraser, simple testimony can be understood as “cases in which a speaker assertorically utters a single – most often, very simple – sentence” such as “Elvis Presley is alive” or “the President is in New York” (Fraser Reference Fraser2021: 4025–26). Simple testimony is perhaps best understood and further defined in contrast to narrative testimony, a category of testimony also introduced and defined by Fraser.

Narrative testimony primarily differs from simple testimony in that it takes on the paradigmatic features of a story (Reference Fraser2021: 4027). Paradigmatic stories are those that describe the doings of humans or human-like characters, taking place over a period of time, encoding some goal structure, and recounting some obstacle to the achievement of aims (Reference Fraser2021: 4027). Not every instance of narrative testimony will include all of these paradigmatic features, as they are not a list of necessary and sufficient conditions, but these are the features that demarcate what Fraser means by a story.

Narrative testimony is also characterized by being a structured and interlocking set of claims, irreducible to its individual parts. To assume that narratives are simply strings of simple testimony, or are reducible to a string of individual sentences is to mistake both the nature of a narrative and the process of making sense of one. On understanding narrative testimony, Fraser writes, “discourse interpretation cannot be thought of simply as ‘scaled up’ sentence-interpretation. Rather, it involves distinctive dynamic processes” (Reference Fraser2021: 4031). This interlocking nature of narrative testimony is notably absent from simple testimony, which instead takes the form of standalone, assertorical sentences.

Further, a key trait of narrative testimony is that it embeds a perspective, which Fraser breaks down into three dispositions: attentive, inquisitive, and interpretive. In terms of the attentive, this disposition asks, what will strike one as interesting or noteworthy? What details will draw them in, and what details will fade into the background or be irrelevant? In terms of the inquisitive, this disposition asks, what kinds of inquiry will be worthwhile? What sorts of investigative methods will one be include to use, and which kinds of explanations will close the inquiry at hand? Finally, in terms of the interpretive disposition, the following questions will emerge: what inferences should one draw, and what evaluations will one be inclined toward? (Reference Fraser2021: 4028). Collectively, these dispositions interlock to form a perspective, and that perspective is embedded in narrative testimony, which then enables testimony to be used as a tool for achieving perspectival coordination.

When interlocutors achieve perspectival coordination, they are able to share a way of looking at the world (4028). Fraser describes this coordination as the audience not just adopting the opinions expressed by the narrative, but structuring and organizing information as the narrative suggests. In this way, perspectival coordination begets perspectival dependence. When the audience structures and organizes information as cued by the narrative, they become “thickly dependent on the testifier” (4047). Fraser argues that the epistemic dependence that emerges through the uptake of narrative testimony is not present with simple testimony.

These perspectival features are hallmarks of narrative testimony, and also help to further map out simple testimony as lacking these features. The simple, assertorical sentences that constitute simple testimony will not embed a perspective, nor will they require perspectival coordination in the way that narrative testimony does. Simple testimony may convey information, and perform other functions of testimonial exchange, but it will not have the rich and complex features that define narrative testimony.

A final feature of narrative testimony is its strong capacity to trigger affective responses. Because narrative testimony embeds a perspective, it encourages the audience to engage first-personally with an experientially rich simulation. Fraser cashes this out in terms of representational formats and argues that narratives can determine what kind of reactive stance ought to be adopted given the interpretive, inquisitive, and attentive dispositions of said testimony (4036–37).

Fraser's rich account of narrative testimony then serves as a foil to simple testimony. The key features of narrative testimony – the interlocking nature, the embedded perspective, perspectival coordination and dependence, the triggering of affective responses – are not present in simple testimony. A simple phrase like “Elvis Presley is dead” doesn't meet the criteria for narrative testimony and so lacks the corresponding epistemic features. I identify testimonial compression, then, as cases where an audience enforces the format of simple testimony upon a testifier.

Consider the following toy example of testimonial compression, an example that I suspect many individuals have experienced. Suppose you have been asked a question that, in your eyes, requires some explanation to answer. Your response is complex, and one that you believe requires contextual details to fully understand. You try to explain and describe the situation, including a timeline of events, motives and reasons behind various actions. To your frustration, your audience interrupts you repeatedly, and insists that you just answer their question with a simple “yes” or “no.” It becomes clear that any other kind of answer will not be accepted. This is an instance of testimonial compression. Through this discourse, the testifier is limited to the response format of simple testimony: “yes” or “no.” Further, this response format is enforced by the audience, who clearly indicates that they will not listen to, or give uptake to, narrative testimony response types – audience uptake is predicated on your testimony's lack of narrative features.

In the following sections, I analyze several examples of testimonial compression that all display different dimensions of testimonial compression. Further, my examples are all from healthcare settings as there is a rich literature on patient-provider communication and medical discourse that helpfully offers many examples, but I don't take testimonial compression to be limited to healthcare or medical contexts. As demonstrated in my toy example, testimonial compression is a common phenomenon, though perhaps frequently utilized in healthcare settings.

2.2. Examples of testimonial compression

In Survey, we see an instance of testimonial compression enacted through a medical form – an audience requests testimonial knowledge from the testifier and patient, Claire. What's more, the possible ways in which Claire can respond to the prompt answers are pre-determined for her: “as much as I ever did; not quite as much now; not so much now; not at all.” In addition to her response options being limited to these options, the enforced format is simple. They are single, assertorical phrases, as Fraser describes, and Claire cannot expand on these options, nor can she wander from them. The audience, through the medical form, has enforced both a format of simple testimony and the available response options. Given this enforcement, and Claire's inability to negotiate these options or respond otherwise, Survey is an instance of testimonial compression.

Note that some prompts or inquiries have format norms that are deeply internalized and socially expected. For instance, if a medical form were to ask for an adult patient's age, it would go against standard norms for the patient to respond in any unit other than years. It would be deeply inappropriate, and frankly annoying without some strong justification, for a patient to give a narrative response to the question of age. Some questions or prompts are particularly well-suited to simple answers, and enforcing a response format (i.e. years, in number form) would not necessarily be compressive because any other way of answering the question would go against firmly established norms. Given this, some kinds of information solicitation have a deeply embedded norm for response formats, and to stray from those norms would be strange. In the instance of asking for age, there are infinite ways in which a person might accurately respond, but the norm is so deeply accepted that anything other than the unit of years and the format of whole numbers would be inappropriate, and options for responses outside this set standard would likely have little value. However, I suspect that such strongly established norms around response formats for the infinite range of inquiries that one could make are in fact quite rare. In Survey, the prompts have no such expectations in terms of response format – “I have looked forward with enjoyment to things” could responded to in myriad ways, from a “yes/no” to a long, winding narrative. This particular inquiry has no obviously appropriate response format, unlike an inquiry about age.

Testimonial compression can be done by way of forms, as demonstrated in Survey, or through a verbal exchange between speakers. Here, I give an instance of spoken testimonial compression used in a medical setting. The following dialogue is from the 2020 film Never Rarely Sometimes Always, written and directed by Eliza Hittman (Reference Hittman2020). To contextualize this exchange, the dialog takes place in an abortion center, where the patient is a young teen from Pennsylvania who has traveled to New York City for an abortion that would be illegal in her home state. She speaks with a counselor at the center, who screens her for intimate partner violence.

Here, the patient's testimony is compressed in real-time, during the conversation that she has with her doctor. The physician, or audience, has set the response options for the testifier, limiting her responses and their simple format. Note that in this verbal exchange, testimonial compression is more softly enforced than the compression in Survey. In the process of filling out forms, a testifier is quite literally unable to negotiate the format of responses, and in some cases the response options themselves. Further, the enforcement of compression is more rigid with digital forms than with paper forms. While a testifier could be more of an anarchist with a paper form, and scribble in margins or respond in ways that go against the form's suggestion, digital forms not negotiable in format.

In Pregnancy, testimonial compression is directed from the audience to the speaker, but its enforcement is significantly less strong than the enforcement found in digital forms. The patient has more space to stray from the proposed response format, and in fact she questions the prompts themselves. She maintains the freedom to deviate slightly from the set responses, but generally chooses to accept the format of questioning. Given this, Pregnancy can be seen as an instance of more negotiable testimonial compression, where a digital form is an instance of less negotiable compression.

2.3. Enforcement and resistance

The strength or rigidity of the enforcement of testimonial compression will vary across contexts and be sensitive to the goals of the exchange, the dispositions and personalities of the interlocutors, further contextual norms, etc. However, if there is no enforcement whatsoever of the testifier's response format, the exchange will cease to involve testimonial compression. The simple response format must be enforced in order to be an instance of testimonial compression.

Relatedly, while I require enforcement of simple response formats, I do not require resistance for testimonial compression to emerge. This is in part simply because a testifier will not always have reason to resist compression. As I noted earlier, some inquiries are well-suited for simple responses, like asking what a person's age is. I suspect that testimonial compression will generally be present but not perceived when the inquiry and enforced response format are well coordinated, similarly to how good design is often not noticed. When things work well, the mechanics and choices that ground good functioning are often not perceived as they create a frictionless experience. Similarly with testimonial compression, there will not always be a reason to resist an enforced simple response format. However, as Theo points out in Survey, I suspect that compression becomes more visible when there is more dissonance, or room for different options, between an inquiry and the enforced response format. When Theo suggests that an inquiry about whether or not a post-partum patient still finds joy in their day-to-day life should be short-answer rather than multiple choice, part of what he is pointing to is a mismatch between the inquiry and the enforced response format. So while the enforcement of testimonial compression may be resisted by the testifier, resistance is not necessary. So long as a simple response format is enforced, testimonial compression occurs.

2.4. Discrete compression

While Survey and Pregnancy are examples of testimonial compression where the compression itself is visible to the testifier, it can also be enacted more discretely, without testifier knowledge or engagement. For example, consider the following transcript of a conversation.Footnote ¹

The compression in Documentation differs from Survey in its relation and visibility to the testifier. In being asked a question with no limitations on response format, the patient isn't restricted in how they answer and they offer a detailed response. Unlike Survey, where the compression is enforced prior to any particular patient filling out a form, or Pregnancy, where the compression happens during a conversation, compression here occurs between the doctor's uptake of testimony and his documentation of what the patient said. Further, this kind of compression is done for an audience external to the two interlocutors. The audience of this medical documentation isn't the patient or testifier, but rather other clinicians that may be involved in the patient's care and need specific pieces of information presented clearly and concisely.Footnote ²

This case may, at first glance, appear to challenge the criteria for enforcement and negates the possibility of resistance. The presence of enforcement is less transparent to the testifier in that the compression is occurring between the physician's uptake of the patient's testimony and its documentation, rather than the compression occurring during the interlocutors' discourse. Unlike other examples, the audience here compresses the testifier's testimony for a different audience of the patient's testimony: other readers of the patient's medical chart. However, whether or not a testifier is aware of the enforcement of simple response formats doesn't determine whether or not they are in fact enforced. So long as this is the case, the example matches the description of testimonial compression. In terms of resistance, the testifier is unable to resist compression if they aren't aware that it is happening. However, enforcement is a necessary criteria for compression, where resistance isn't. So while compression may be more or less evident to the testifier, testifier knowledge of compression isn't necessary for a case to count.

4.1. Testimonial compression and information extraction

Sometimes, the goal of testimonial exchange is primarily information extraction or knowledge gathering. People are asked to fill out surveys all the time, for a variety of purposes. Surveys are useful tools for the exchange of testimonial knowledge in that they are efficient, pre-emptively organize the knowledge being solicited, and as a result make the information collected reusable for later purposes. For example, in medical contexts, testimonial compression via medical forms may make patient responses useful for future analysis and research, or easier to process into external databases.

The streamlining that compression enables can also be extremely useful for reducing informational lag, or the time that it takes to transfer information from one place to another. It can also be instrumental in making information more digestible, or more useful for a particular audience. For instance, the specific format and structure of medical notes medical charting enhances the utility of that information. Medical charts are a keystone in hospital communication, and by enforcing a specific set of expectations for how medical information is documented, the audience can more easily find the information they need (Hobbs Reference Hobbs and Iedema2007: 41).

Testimonial compression for the purposes of information extraction can also be useful for triggering further inquiry. In Survey, Theo ruminates on how post-partum depression can't be accurately captured in the offered response options. While he may be correct, it is possible that the survey is intended merely to flag patients as potentially experiencing post-partum depression rather than as fully supplanting the diagnostic process. If Claire were to respond in a way that indicated in no way that she was struggling, her doctor may not feel the need to follow up on the subject. However, were Claire to respond with only “d. Not at all,” her physician may then have reason to ask her more about it. So while Theo is skeptical of the form's diagnostic capabilities, it may in fact serve a lesser purpose than that.

Testimonial compression, when used for efficient information extraction, can also vary in its impact on the emotional valence of an interaction. In many cases, information extraction can feel cold and impersonal (Armstrong et al. Reference Armstrong, James and Dixon-Woods2012; Dichter et al. Reference Dichter, Makaroun, Tuepker, True, Montgomery and Iverson2020; Lorimer et al. Reference Lorimer, Cox and Langford2012; Skinner et al. Reference Skinner, Allen, McIntosh and Palmer1985). When information extraction is cold and extractive feeling, patients report the downstream effect of feeling reduced to their illness, or their illness experience as broken into fragments of information (Akther et al. Reference Akther, Molyneaux, Stuart, Johnson, Simpson and Oram2019; Todres et al. Reference Todres, Galvin and Dahlberg2014).

A further downstream effect of information extraction done in ways that are cold, impersonal, or brusque can include leaving patients feeling uncared for. The provision of care has many different aspects, and getting good information and medical histories from patients is undoubtedly important (Hassan Reference Hassan2018; Ratna Reference Ratna2019). However, communication that conveys care for the patient can be equally important (Jangland et al. Reference Jangland, Gunningberg and Carlsson2009; Markides Reference Markides2011; Vermeir et al. Reference Vermeir, Vandijck, Degroote, Peleman, Verhaeghe, Mortier, Hallaert, Van Daele, Buylaert and Vogelaers2015; Warnecke Reference Warnecke2014). Testimonial compression, then, can run the risk of being used in a way the optimizes information extraction at the expense of caring communication.

Testimonial compression, if used in a way that disrupts quality communication, can also inhibit some of the benefits that patients experience from being listened to in medical contexts. There is growing evidence that physicians listening to their patients has, at minimum, instrumental value in clinical settings. Fisher, from Emergency, acknowledges this himself – “they all want to tell their story…people queue as though they're waiting to meet an oracle” (Fisher Reference Fisher2022: 68–69). Meghan O'Rourke, author of The Invisible Kingdom: Reimagining Chronic Illness, discusses this in further detail. In chronicling her years-long struggle with chronic illnesses, she writes,

O'Rourke points out a similar effect in a study on patients with irritable bowel syndrome (IBS). These patients were told that they were participating in a study examining the benefits of acupuncture. One group was assigned practitioners who only had “perfunctory” and “brusque” interactions with patients. The second group was treated by a researcher who “warmly asked questions about them and expressed empathy for their suffering.” Both groups were given identical sham treatments, and the symptom reduction in the second group was higher than the first. What's more, the percentage of patients in the “empathetic” group that reported adequate relief in symptoms was as high as those who reported symptom relief in clinical trials of drugs prescribed for IBS (O'Rourke Reference O'Rourke2022: 67–68). Other studies have been done that document similar effects, indicating that more empathetic and high-quality communication alone can benefit patients (Egbert et al. Reference Egbert, Battit, Welch and Bartlett1964; Fuentes et al. Reference Fuentes, Armijo-Olivo, Funabashi, Miciak, Dick, Warren, Rashiq, Magee and Gross2014; Hojat et al. Reference Hojat, Gonnella, Nasca, Mangione, Vergare and Magee2002, Reference Hojat, Louis, Markham, Wender, Rabinowitz and Gonnella2011; Kaptchuk et al. Reference Kaptchuk, Friedlander, Kelley, Sanchez, Kokkotou, Singer, Kowalczykowski, Miller, Kirsch and Lembo2010; Ofri Reference Ofri2017: 69–75; Rakel et al. Reference Rakel, Barrett, Zhang, Hoeft, Chewning, Marchand and Scheder2011; Rao et al. Reference Rao, Anderson, Inui and Frankel2007; Street Jr. et al. Reference Street, Makoul, Arora and Epstein2009).

As demonstrated in Emergency, testimonial compression can be used in a way that paints an interaction as “perfunctory” and “brusque,” rather than empathetic or caring. By definition, narrative compression reduces patient narratives, and the patient's own perspective is often lost or washed away in this process. Cases where patients experience and resist narrative compression, then, seem likely to miss the mark in terms of producing the benefits that high-quality communication can have.

Testimonial compression may also interfere with patients disclosing medically relevant information. Because testimonial compression (as I have introduced it) is a novel concept, there is no data to support this hypothesis. However, there is significant data that poor communication can increase the risk for misdiagnosis (Muhrer Reference Muhrer2021) and medical errors (Frydenberg and Brekke Reference Frydenberg and Brekke2012; Lazris et al. Reference Lazris, Roth, Haskell and James2021; Murphy and Dunn Reference Murphy and Dunn2010; Risser et al. Reference Risser, Rice, Salisbury, Simon, Jay and Berns1999; Shitu et al. Reference Shitu, Hassan, Aung, Kamaruzaman and Musa2018). There is also significant evidence, in the opposite direction, that effective communication improves clinical decision-making and quality of care (Drossman Reference Drossman2013; Drossman and Ruddy Reference Drossman and Ruddy2020; Gaddis Reference Gaddis2019; Patak et al. Reference Patak, Wilson-Stronks, Costello, Kleinpell, Henneman, Person and Happ2009). By extension, these studies support the hypothesis that when compression disrupts patient-provider communication, or lowers its quality and effectiveness, it can also increase the risk for misdiagnosis or medical errors. Cases of testimonial compression that pre-emptively decide for the patient what is medically relevant and what isn't could lead to missed information that wasn't solicited from the patient.

However, information extraction by way of testimonial compression need not have these negative effects. Compression is a tool, and it can be used to benefit patients, as evidenced in Pregnancy. This case demonstrates an instance where testimonial compression is used as a strategy to solicit highly sensitive, health-related information from a distressed patient without requiring them to disclose any more than they need to. The nuances of this interaction aren't well-captured in the transcript, but the scene in the film depicted a nuanced and careful use of testimonial compression. For one, the clinician posing questions and directing the patient to answer them with the limited “never/rarely/sometimes/always” response options isn't depicted as cold and extractive. In fact, her soliciting of information from the patient in this way reads as a form of care – it allows the patient to provide the information needed, but only the information needed, in recognition that providing this kind of deeply personal information to a stranger can be distressing, unusual, and challenging.Footnote ³ Further, in limiting the patient's response options to her questions, the clinician also helps the patient maintain some privacy.

4.2. Testimonial compression and epistemic harms

As I have discussed, the use of testimonial compression need not be harmful. However, testimonial compression can be used to enact the specific epistemic harms of participatory and informational prejudice toward patients.

Ian Kidd and Havi Carel have extensively discussed epistemic harms within healthcare (Carel and Kidd Reference Carel and Kidd2014, Reference Carel and Kidd2017; Kidd and Carel Reference Kidd and Carel2017, Reference Kidd, Carel, Barker, Crerar and Goetze2018, Reference Kidd and Carel2019). They argue that some of the kinds of epistemic injustices that patients experience result what Gail Polhaus Jr. calls truncated subjectivity (Pohlhaus Jr. Reference Pohlhaus2014). Polhaus, Jr. writes that in cases of truncated subjectivity, the testifier's “epistemic labor contributes to the community via which epistemic interests are pursued, but she is not permitted to contribute in ways that would redirect epistemic practices toward those parts of her experienced world” (Reference Pohlhaus2014: 107).

Kidd and Carel argue further that ill persons and patients often experience participatory and informational prejudices. Participatory prejudice occurs “when a person or group is prejudicially judged to lack capacities required for having a sense of relevance, and hence as not being suitable participants for collective epistemic activity” (Kidd and Carel Reference Kidd and Carel2017: 180, emphasis original). Ill persons are particularly vulnerable to participatory prejudice because first, they are assumed to “lack the training and experience needed for the possession of a robust sense of relevance…to make meaningful contributions to the epistemic practices of medicine” and second, because ill persons are “typically regarded as the objects of the epistemic practices of medicine rather than as participants in them” (180–81).

Informational prejudice “occurs when a person or group is prejudicially judged to lack the ability to provide information relevant in a given context and hence as being an unsuitable participant in collective epistemic activity” (Reference Kidd and Carel2017: 181; Hookway Reference Hookway2010). This prejudice can manifest as a refusal to concede the relevance or significance of the information being offered by a testifier, or as a refusal to consider presuppositions about the significance and types of information that are legitimate and admissible in the epistemic project (Kidd and Carel Reference Kidd and Carel2017: 180–81). When ill persons experience informational prejudice, they are subject to these two refusals.

Kidd and Carel's account relates to testimonial compression in that testimonial compression, and more specifically narrative compression, can be a way of enacting these same kinds of epistemic harms. For example, in Emergency the patient's testimonial knowledge is solicited, and she isn't disbelieved about her symptoms despite the doctor's belief that the patient is mistaken about not having Covid. Rather, the patient's testimonial contributions are truncated such that she is only permitted to testify within an assigned format, in response to the doctor's close-ended questions, and her testimony outside of these constraints is ignored. She is able to contribute in the provision of factual information (simple testimony), but not able to describe her distinctive experiences (narrative testimony). In terms of informational and participatory prejudices, the patient's own perspective on her illness is dismissed as irrelevant for the epistemic activity of determining her diagnosis. While testimonial compression is not always harmful, it can be a way of enacting informational and participatory prejudices in the ways the Kidd and Carel articulate.

6.1. Aspects of negotiation

Negotiations of testimonial compression entail the testifier negotiating the enforcement of simple response formats. The identification of this kind of negotiation is compatible with Lazare's claim that sometimes the negotiations are not about the medical problem itself, but about how the medical problem is made known.

Consider again Emergency. Not only is this conversation an example of verbal narrative compression, it is also an example of a negotiation of the patient's testimony's narrative form. Fisher, the physician, attempts to enforce both the kinds of questions answered and the patient's response format (simple), but the patient repeatedly resists. Her insistence on starting over when interrupted transforms the exchange into a negotiation of compression, unlike Pregnancy where the patient doesn't resist the imposed formatting. Emergency resembles many aspects of integrative bargaining and Lazare's mapping of potential sources of conflict – it involves interlocutors having a shared goal, the care of the patient. However, conflict emerges over the nature of the patient's problem. The concept of testimonial compression enables me to identify more specifically a negotiation over how the patient is able to convey knowledge about her problem, which shapes Fisher's identification and understanding of it.

Emergency also resembles some aspects of distributive bargaining. Fisher and his patient do not start with equal levels of presumed epistemic authority – Fisher obviously holds an advantage. Further, Fisher is beholden to many more patients and potential negotiations, while the patient has no such ties to other patients. This leaves the patient not only vulnerable in that she needs medical care, disadvantaged in that Fisher serves as a gatekeeper to further care, but also in that she must compete for her physician's extremely limited time. Both epistemic authority and time are negotiated resources in this case. All of these aspects have a cumulative effect on the patient's ability to resist Fisher's compression of her testimony from narrative to simple.

6.2. Negotiating medical relevance

Testimonial compression can also be negotiated in the way that it is used to draw the parameters of relevance. As W. A. Rogers describes, the physician is tasked with “filtering” patient testimony through a “medical sieve” (Reference Rogers2002: 79), which by extension enforces parameters of medical relevance.Footnote ⁴ This is further presumed to be necessary because “the patient cannot be trusted to know the relevant details from the insignificant” (79). Rogers's reasoning is undoubtedly true at least some of the time. But as Popowicz argues, patients' medical information and corresponding medical decisions are rarely merely medical (Popowicz Reference Popowicz2021: 10). Further, being overly dismissive of the relevance of information that patients provide can have a negative effect on both the patient-provider relationship and the overall negotiation.

An instance of this can be found in Susan Greenhalgh's autoethnography on her experience with a fibromyalgia diagnosis. Greenhalgh's physician was (allegedly) a fibromyalgia specialist, albeit a controversial one, and treated her for nearly a year. His prescribed treatment was extremely aggressive, requiring her to eventually refrain from all physical activity or exercise, take up an extremely restrictive diet, and cease work as an academic. After months of suffering and personally documenting her decline, Greenhalgh confronted her doctor and accused him of both falsely diagnosing her and of dismissing her documented evidence that her symptoms were not improving. She writes,

This exchange helps demonstrate two key things. First, that in her physician's eyes, Greenhalgh's whole illness experience was compressed into physiological measurements, not just her testimony about it. Second, the conversation demonstrates a negotiation of relevance. Dr D doesn't necessarily indicate that he disbelieves Greenhalgh's detailed documentation of her steady decline. The conflict isn't a matter of her trustworthiness as a testifier.Footnote ⁵ Rather, he demonstrates that her testimony, regardless of its truth, is simply irrelevant. The scope of medical relevance is narrowly defined as quantitative, and as detached from Greenhalgh's daily experience of illness.Footnote ⁶

This negotiation also fails in an important way. Borrowing from the integrative bargaining model, there is no achievement of a shared goal. Further, neither party leaves the interaction giving much epistemic weight to the other's opinion or beliefs – they more or less end where they began, epistemically speaking. The core issue, Greenhalgh's medical situation, is unresolved and Greenhalgh feels disempowered as a patient.

While the identification and discussion of negotiations in medical discourse is nothing new, testimonial compression is a kind of negotiation that I add to the existing taxonomy. When negotiating compression, patients negotiate how it is that their medical problems are made known, the relevance of their perspective or narratives, or some combination of these. Further, different types of testimonial compression will be more or less amenable to these kinds of negotiations.