Are Metaphors Ethically Bad Epistemic Practice? Epistemic Injustice at the Intersections

Abstract

The COVID-19 pandemic brought the debate about the ethics of metaphors to the fore. In this article, I draw on blending theory—a theory of cognition—and theories of epistemic injustice to explore both the epistemic and ethical implications of metaphors. Beginning with a discussion of the conceptual alterations that may result from the use of metaphors, I argue that the effects these alterations have on available hermeneutical resources have the potential to result in a type of hermeneutical injustice distinct from the “lacuna” described by Miranda Fricker (Fricker 2007). Following, I examine how metaphors may therefore be considered “ethically bad epistemic practice,” as described by Rebecca Mason, because of how they may contribute to perpetuating an inequitable epistemic status quo (Mason 2011). Yet these same features may be used to promote epistemic justice in the context of intersectional power relationships. Situating the effects of metaphors within an inequitable yet dynamic epistemic system, I argue that foregrounding intersectional power dynamics enables us to interrogate the ethics of metaphors with consideration of both the epistemic and material consequences that may occur. I conclude by providing guidance for how, given that metaphors do epistemic work, we may use them to do ethical epistemic work.

In 2020, at the height of the COVID-19 pandemic, the #ReframeCovid initiative sought contributions of non-war-related figurative language to promote alternative ways to understand the global pandemic (The Initiative 2020). The call went out for people all over the world to contribute to an open-source collection of visual and linguistic metaphors to provide alternatives to the war metaphors that dominated how people around the world framed, and therefore responded to, the pandemic (Gök and Kara 2021; Vlastou 2021). Criticism arose about how war metaphors justified the continued marginalization of women's health (Bailey, Shankar, and Phillips 2021), supported the agenda of those with political power (Chapman and Miller 2020), and fueled xenophobia (Khan, Iwai, and DasGupta 2021). If metaphors are to be blamed for playing a role in perpetuating these social problems to the extent that a global movement arose to change them, it is probable that metaphors have both epistemic and ethical consequences. At present, however, we lack a framework to explain how and why this is the case.

In this article, I draw on blending theory—a theory of cognition—and theories of epistemic injustice to map an explanatory framework that reveals both the epistemic and ethical implications of metaphors. First, I outline blending theory and the relationship between metaphors and their potential epistemic consequences. Second, I explain two ways that metaphors shape our hermeneutical resources and how these have the potential to contribute to epistemic injustice. For each, I outline how metaphors that lead to epistemic injustice by contributing to the inequitable distribution of epistemic resources, and thereby supporting an inequitable status quo, may constitute ethically bad epistemic practices.

In the following section, I examine the conceptual impacts of metaphors in the context of the multidirectional power relationships and diverse identities characteristic of lived experience. I consider how metaphors used to support social change create ethical tensions that distribute epistemic resources in ways that can both support and undermine efforts toward epistemic justice. What distinguishes those that are ethical from those that are not is their relationship to the dominant discourse and their impact on concepts that capture group experience and influence credibility judgments. These considerations ought to be contextualized within an inequitable epistemic system. Then I address the debate around the use of the war metaphor in healthcare contexts. I suggest that the considerations I present provide resources to help determine whether a metaphor may constitute ethically bad knowledge practice, particularly considering the complexity of intersectional power relations within inequitable epistemic systems.

I. Metaphors and Blending Theory

“Blending theory” emerged from the cognitive sciences (Fauconnier and Turner 1998). It has since influenced research in mathematics, social science, literature, and linguistics. Blending theory is a theory of mental operation that explains the construction of meaning (Fauconnier and Turner 2003). In this theory, mental spaces are “small conceptual packets constructed as we think and talk, for purposes of local understanding and action” (Fauconnier and Turner 2003, 58). These spaces are dynamic and can be modified as discourse unfolds (Fauconnier and Turner 2003). Within this theory are multiple types of conceptual integration. The type of integration relevant to understanding metaphors is double-scope (Fauconnier and Turner 2003). Double scopes create a frame of reference between two fundamentally different, clashing inputs (Fauconnier and Turner 2003), such as the two concepts that are brought together to make meaning with a metaphor. Metaphors, according to blending theory, fuse concepts in a way that prompts the construction of mental spaces that generate meaning (Fauconnier and Turner 1998). Metaphors do this by blending what is known about each concept used in the metaphor—for example, the input concepts of “men” and “wolves” in the metaphor “men are wolves”—in the context of generic background knowledge, to generate new meaning (Fauconnier and Turner 1998; Hart 2008). This emergent meaning is the product of three steps: 1) composition: the fusion of the domains—the input spaces—and the ideas associated with them, 2) completion—the integration of select background information with the fused ideas, and 3) elaboration—the cognitive work that emerges from the generated frame of reference, such as reasoning and drawing inferences along the lines of the metaphor (Fauconnier and Turner 1998; Hart 2008). For example, in the metaphor “men are wolves” we may consider the steps: 1) composition: fusing the ideas of men and wolves; 2) completion: integrating select background information, such as narratives involving anthropomorphized wolves from the fable “Little Red Riding Hood,” or a wolf in sheep's clothing, and the context of the metaphor to create a framework for understanding what is being expressed; and 3) elaboration: reasoning about how to act as a man if men are conceptualized as wolves, such as acting with cunning or violence, or how to interact with men conceptualized as a wolves, such as being suspicious or guarded against hidden violent intentions.

The relationship between the ideas fused makes some background information more likely to be recruited than other information. In the example above, not all information known about wolves or men is likely to be recruited through the metaphor comparing men to wolves, such as wolves traveling in packs, howling at the moon, or running fast, because these are less salient in the fusion of the ideas “men” and “wolves,” given the background context. This partial recruitment contributes to metaphors’ ability to both create and persuade, something that has made them “philosophy's greatest enemy and also its greatest ally” (Ricœur 2003, 10). The result is that metaphors evoke a frame of reference that influences reasoning in a way that is both partial and evaluative (Deignan 2010). When only select generic background information is recruited, “speakers may choose to recruit [a] particular structure in order to promote a certain perceived reality” (Hart 2008, 97). This select information reinforces a particular ideology (Hart 2008). One ideology reinforced by the above example metaphor may be heteronormative masculinity, emphasizing strength, power, cunning, and violence (Manne 2018).

In the elaboration stage, the metaphor “becomes reality” as it is entrenched in patterns of thought and action that convey and reinforce particular ideologies (Hart 2008). When the relationship between “men” and “wolves” becomes the framework within which one reasons about what “men” are, one is then likely to act based on this reasoning (Lakoff and Johnson 2003). Studies examining the implications of metaphors have found that people are more likely to endorse actions that align with the metaphorical framing of a problem than ones that conflict, such as responding with force when crime is metaphorically described as a beast (Thibodeau 2016) or using antidepressants if mental illness is framed as a chemical imbalance (Kemp, Lickel, and Deacon 2014). Through metaphors, “the audience is persuaded to adopt specific opinions which reflect the ideology of those who have the power to create that discourse and disseminate their metaphors” (Efeoğlu Özcan 2022, 171). In addition, because “people in power get to impose their metaphors” (Lakoff and Johnson 1980), the reality conveyed through dominant metaphors reflects the experience of those in positions of social power and the ideologies that support their status. Used frequently enough, the knowledge produced through elaboration may shape the background knowledge recruited for future understanding (Fraser 2018). For example, frequent use of the “men are wolves” metaphor may contribute to what background information is recruited when the ideas of “men” and “wolves” are used in future discourse. When the reality the metaphor conveys reflects the experience of those empowered by the dominant discourse, these experiences are validated, such as displaying or accepting heteronormative masculinity; experiences of reality that contradict or are excluded by the metaphor are obscured.

Although metaphors are traditionally defined as “understanding and experiencing one kind of thing or experience in terms of another” (Lakoff and Johnson, 1980, 455), when viewed in the context of blending theory (Fauconnier and Turner 1998; Hart 2008), it is more appropriate to define a metaphor as evoking a frame of reference within which one understands and experiences one or both things in relation to a partial understanding of the other. What this definition adds is the recruitment of select, partial background information for generating new meaning (the completion stage described above) and the production of not only an idea or experience, but one that may then be generative of additional reasoning (the elaboration stage described above). It also allows for nontraditional metaphors, such as models (Black 1960; Brown 2003), archetypes (Black 1960), and analogy (Childress 2012), to be included within this definition, and thus be analyzed through this lens. In addition, by acknowledging metaphors’ conceptual generative potential, it acknowledges that ideas contributing to the metaphor may be altered by its composition. This alteration of conceptual space has both epistemic and ethical consequences as it influences both understanding and subsequent behavior. Defining metaphors in this way shifts the conversation from what metaphors “are” to what metaphors “do.” These components acknowledge the possibility that metaphors may do diverse epistemic work within dynamic knowledge systems.

II. Metaphors as Ethically Bad Knowledge Practices

Using this model to understand how metaphors work highlights the inherent partiality of the frameworks for reasoning that they produce. Partial recruitment of the ideas associated with each concept in the context of relevant background information blends to produce a framework that leaves out some conceptual elements of each concept. The elaboration that subsequently occurs neglects consideration of these left-out conceptual elements. I argue this feature gives metaphors the potential to do at least the two following types of potentially harmful epistemic work:

1. 1. Metaphors can direct us away from, and distort, our understanding of the concepts the metaphors are otherwise illuminating.

2. 2. Metaphors can exploit the terms they use to elucidate a concept or experience in ways that distort one or both terms, so they are no longer useful for their main purpose.

In both cases, metaphors may affect the epistemic resources available in ways that can make them less useful for understanding or conveying experience, causing epistemic harm.

For example, a metaphor that causes epistemic harm by altering the concept it intends to illuminate may be found in the metaphor “the body attacks itself” to describe the experience of autoimmune disease. Beth Ferri discusses her experience with autoimmune disease, arguing that although the concept this metaphor evokes does capture some aspects of her experience, this metaphor casts her as both the enemy and the hero, which does not acknowledge important parts of her experience, including the interconnection between the immune system and other aspects of her body and everyday life (Ferri 2018). The metaphor “the body attacks itself” excludes these elements from the concept of “experiencing autoimmune disease,” reducing the concept and making it more challenging for Ferri to convey her experience. By altering a concept so that it no longer captures important components of experience that “experiencing autoimmune disease” otherwise may have, the epistemic resources to understand and communicate this experience are obscured.

In this case, the frame of reference produced through the metaphor excludes elements of the concept it seeks to elucidate, altering the epistemic resources available for conveying experience. However, metaphors may also cause epistemic harm by exploiting the terms they use to convey an experience so that they are no longer useful for their main purpose. For example, some concepts accurately capture experiences because of their strength, such as misogyny (Manne 2018), and rape (Fraser 2018). When these words are used as the input domain in metaphors that draw comparisons between this domain and others that weaken the concept's ability to capture the significance of the experience, the result is epistemic harm because members of particular groups are no longer able to accurately communicate their experience when it is in their interest to do so (Fraser 2018). As such, using the metaphor “rape the fields” when talking about crops, or comparing winning a sporting event to “raping” the other team, trivializes the language concept of rape in a way that takes away the power of the language to communicate the actual experience of rape (Fraser 2018). The metaphor diminishes the epistemic resources of those who have an interest in communicating their experience. Although the experience may still be understood, the epistemic resources to communicate it are rendered insufficient. They are distorted because they are altered by their involvement in metaphorically conveying a different experience.

III. Altering Concepts as Epistemic Harm

Metaphors can therefore cause epistemic harm by either altering the concept they wish to illuminate or altering the concepts they use in ways that negatively affect the ability of certain groups of people to convey their experience. I argue this constitutes epistemic injustice, “a wrong done to someone specifically in their capacity as a knower” (Fricker 2007, 1), because it deprives groups of people of the existing epistemic resources to understand and convey their experience.

I argue that these conceptual alterations that metaphors may produce can take two related forms: exclusion (leading to reduction) and dilution. In the case of exclusion, elements of a concept are “pushed out” of the conceptualization so that they are not captured by the frame of reference produced. The “body attacking itself” metaphor is an example of this. Elements of experience that may otherwise have been associated with autoimmune disease, such as a lack of dichotomy between the “body” and the “self,” cannot make sense within this framing (Ferri 2018). What otherwise may have received conceptual space is excluded. This exclusion may result in a type of conceptual foreclosure by prematurely limiting alternative conceptualization, leading to epistemic oppression (Dotson 2012).

In contrast, a metaphor may distort a concept by bringing in additional conceptual elements, directing attention away from those that are key components of experience. The above metaphors using “rape” are examples. In these cases, additional elements are being added to the concept— losing a game, or farming—that distort the concept so that it no longer has the same power to convey the experience for which it was originally intended. These metaphors still leave space for the original meaning of the concept but dilute it by focusing attention away from this meaning.

These two conceptual alterations are distinct, but they are related in that they both result in the distortion of concepts in ways that prevent the use of these concepts from adequately conveying experiences. However, distortion by dilution still leaves room for the experience to be captured within the frame of references (albeit less saliently than it otherwise would), whereas distortion by exclusion prevents the concept from capturing experiences for which it was otherwise intended.

The epistemic injustices produced through metaphors in these ways are a type of hermeneutical injustice distinct from the lacunae in resources that Fricker describes as contributing to hermeneutical injustice. Fricker argues that hermeneutical injustice occurs when someone “has a significant area of their social experience obscured from understanding owing to prejudicial flaws in shared resources for social interpretation” (Fricker 2007, 147). This is arguably what occurs in the above-described cases; the alterations to either the concepts used or illuminated by the metaphors dilute or exclude conceptual elements in ways that distort these concepts, creating flaws in these shared conceptual resources for social interpretation. When metaphors work in this way, we may therefore say that they result in hermeneutical injustice. However, this injustice is not the result of a “a collective hermeneutical lacuna” (157) where prejudicial flaws in shared epistemic resources have left a yet-to-be-filled gap that prevents one from rendering their experience intelligible. In this case, there is no negative space or gap in concept. Elements have either been pushed out, leaving no space for those needed to convey experience, or are there, but diluted beyond utility. This “hermeneutical pruning” cuts back conceptual resources, shaping and limiting the direction within which future hermeneutical growth is possible, and “hermeneutical dilution” decreases the conceptual saliency and therefore the utility of the concept.

Instead, these types of injustice are like the second type of unknowing described by Rebecca Mason in her critique of Fricker's explanation of hermeneutical injustice. Mason argues that the “gap” Fricker describes “is ambiguous between two kinds of unknowing: an unknowing to which members of non-dominant social groups are subject by virtue of their systematic hermeneutical marginalization and an unknowing to which members of dominant groups are subject by virtue of their ethically bad knowledge practices” (Mason 2011, 295).

Mason suggests that ethically bad knowledge practices are those that “maintain gaps in dominant hermeneutical resources even while alternative interpretations are in fact offered by non-dominant discourses” (301). They contribute to miscognition, an epistemic agreement to see the world wrongly, sustaining false beliefs (Mills 2007; Mason 2011). Drawing on Charles Mills, Mason argues that the ignorance produced is “a kind of epistemically culpable and morally noxious miscognition that facilitates the maintenance of the status quo” (Mason 2011, 302). Both types of conceptual alterations resulting from the work that metaphors do may contribute to the type of miscognition that Mason describes. However, as with Fricker's definition on which Mason draws, “maintaining a gap” is insufficient to capture the type of miscognition that metaphors may facilitate. I therefore argue that epistemically bad knowledge practices also include those that prune or dilute conceptual elements from dominant hermeneutical resources when the available dominant hermeneutical resources would otherwise have been sufficient to convey experiences of those who are epistemically marginalized. Metaphors may therefore produce ignorance through promoting hermeneutical injustice by taking away hermeneutical resources useful for conveying experience. Through these means, metaphors can facilitate miscognition that maintains or exacerbates the status quo, which can be considered ethically bad epistemic practice.

IV. Conceptual Resources and Testimonial Injustice

I have argued thus far that the conceptual alterations that result from metaphors may decrease the utility of the concepts they use and evoke, impairing one's ability to adequately understand and communicate experience, and that this constitutes a type of hermeneutical injustice. However, due to the relationship between the available hermeneutical resources and testimonial injustice (Fricker 2007), alterations in concepts that result in dilution and exclusion, when those concepts are meant to capture qualities that impact credibility judgments, may be particularly problematic.

Metaphors may entrench group-based credibility errors by reinforcing stereotypes. Stereotypes can emerge or be reinforced through metaphors (Young et al. 2019). For example, repeated use of metaphors that compare “experiencing dementia” to “being a zombie” create and reinforce a stereotype of dementia that renders those who experience it as inhuman and incompetent (Young et al. 2019). The fusion of ideas associated with zombies, such as “emptiness,” “brainless,” “contagion,” “inhuman,” and “living dead” with ideas associated with “dementia” produces a frame of reference within which reasoning that a person with dementia lacks credibility makes sense because markers of credibility are pushed out of the concept. The result may be that those with dementia self-silence their experience out of fear they are unreliable (Young et al. 2019), what Kristie Dotson refers to as “testimonial smothering” (Dotson 2011). It may also result in those with dementia not being considered knowers and therefore not identified as valuable knowledge-contributors in decisions that are in their interest (Young et al. 2019), what Dotson refers to as “testimonial quieting” (Dotson 2011). The frame of reference this metaphor supports justifies reasoning that a person is “cognitively unreliable, emotionally compromised, existentially unstable or otherwise epistemically unreliable in a way that renders their testimonies and interpretations suspect simply by virtue of their status as an ill person with little sensitivity to their factual condition and state of mind” (Carel and Kidd 2014, 531).

This would constitute ethically bad epistemic practice as the miscognition it produces denies one the credibility resources they are due and prevents potentially valuable testimony from being included in the knowledge-production process. This occurs when the concept “dementia” is pruned by using metaphor to exclude markers of credibility. This entrenches ageist and ableist ideologies.

In addition, testimonial injustice may occur when a metaphor produces or reinforces a stereotype that encourages a group-based credibility error that is a credibility excess, rather than a credibility deficit. For example, the metaphor “just what the doctor ordered” or “doctor's orders” may evoke a frame for reasoning in which a physician is given more credit than they are due based on their status as a physician, rather than their actual credibility on a particular subject. This metaphor, in the Western context, which compares the physician to a “military general” conceptually reinforces a paternalistic and authoritarian relationship between a doctor and patient (Nie et al. 2016). This may bolster physicians’ testimony with a credibility excess because of their status as physicians rather than their epistemic credentials. Additional credibility resources become included in the concept of “physician” when they otherwise may not have been, detracting attention from the fallibility of physicians and the limitations of their knowledge. Because of this excess credibility, those not members of this group may also self-silence or not be consulted when their testimony could contribute to knowledge-production, thereby producing testimonial smothering and quieting (Dotson 2011). Those who do contribute may be ignored because of their comparably lower group-based credibility to that of the physician, thereby producing testimonial injustice (Medina 2011).

In both cases, alterations to the conceptual categories within which people are viewed influence the perceived credibility they are due. Because what is being diluted or excluded are credibility resources that influence credibility judgments, making an accurate judgment less likely, these metaphors facilitate testimonial injustice. When these credibility resources are distributed to grant those in privileged positions excess credibility or create a credibility deficit for those in marginalized positions, the unequal distribution of epistemic resources maintains the status quo, and the metaphor may constitute ethically bad epistemic practice.

Thus far, I have argued that metaphors may facilitate epistemic injustice by producing alterations in either the concepts on which they draw or those they seek to illuminate. These alterations may be exclusions, whereby the concept is pruned to exclude important conceptual elements for capturing experience, or dilutions, whereby the concept is altered to include additional conceptual elements in ways that detract attention from important conceptual elements for capturing experience. When this occurs through metaphors that aim to capture experience, hermeneutical injustice may occur as conceptual resources for conveying an experience are made insufficient. When metaphors alter conceptual categories by diluting or excluding information that is used to make credibility judgments in ways that make them inaccurate, testimonial injustice may occur.

V. Social Structures and Power Dynamics

What has so far received insufficient attention in this analysis is the role of social structures and power dynamics. As noted in the introduction, not all metaphors have the same status within inequitable social epistemic conditions where the dominant discourse is the reality against which “truth” is judged (Lakoff and Johnson 1980): “the people who get to impose their metaphors on a culture get to define what we consider true—absolutely and objectively true” (Lakoff and Johnson 2003, 160). As Mason describes, failure of the dominant hermeneutical resources to adequately convey experience does not preclude the possibility that there are other nondominant hermeneutical resources one can use to convey their experience. As such, altering a concept within the dominant hermeneutical resources does not necessarily mean an alternative conceptualization cannot exist. Rather, injustice may occur because the inequitable epistemic status quo is maintained and/or reinforced through the conceptual distortions that metaphors produce in the dominant hermeneutical resources, where these resources hold the social status of “truth.”

Considering that the status quo of the dominant hermeneutical resources is unjust, maintenance or proliferation of this status quo furthers its injustice (Mason 2011). However, because metaphors can distort conceptual resources, they can also distort concepts as they exist within the dominant hermeneutical resources in ways that make clear experiences that may otherwise be obscured. In this way distortion caused by metaphors can facilitate a truncated form of epistemic justice. Take, for example, the metaphor in which a wheelchair is referred to as a “corvette.” In this case, the metaphor pushes the concept of using a wheelchair to better reflect more aspects of the experience of using a wheelchair; the experience of gaining mobility, freedom, and independence through using a wheelchair may feel similar to getting a car. This “freedom” is counter to a dominant discourse in which a person using a wheelchair is described as “confined” or “trapped.” The metaphor dilutes the concept of “the experience of using a wheelchair” to also include elements of freedom, expanding the concept to reflect a wider variety of experiences, when those experiences have historically been problematically excluded from the dominant discourse. In cases like this, the dilution expands conceptual resources in ways that include realities that may otherwise be excluded, challenging the problematic status quo that the dominant discourse perpetuates.

However, because contradicting or challenging the dominant discourse may be considered a mark of lack of credibility, using metaphors such as this may result in a person receiving an undue credibility deficit. If those who impose their biased metaphors consider the metaphors and the ideologies they support to be absolute truth (Lakoff and Johnson 2003), distorting these metaphors to include hermeneutical resources that are excluded from those that are dominant may be perceived as conveying an untruth. This may result in a credibility deficit that increases the likelihood that one will experience testimonial injustice. Building on the example above, someone who metaphorically frames a wheelchair as a corvette may be dismissed as childlike or delusional for conveying an inaccurate representation of reality, when reality is measured against the dominant discourse. Insofar as markers of credibility within the dominant discourse include communicating using hermeneutical resources that maintain and proliferate the inequitable status quo, distorting concepts in ways that challenge the status quo, while promoting hermeneutical justice, may result in creating testimonial injustice. Checks and balances within the epistemic system work to sustain an inequitable status quo by invalidating the testimony of those who challenge it. This testimonial injustice, however, is the possible result of using the metaphor in a social context; it is not inherent to the work that the metaphor does on a conceptual level. Yet this result has very real impacts when metaphors are used by people within inequitable social systems—impacts with epistemic consequences. This complicates the judgment of whether a metaphor is ethically bad epistemic practice.

Concurrently, because metaphors that support the dominant discourse may be considered a mark of credibility, they may be used to bolster the credibility of someone who may otherwise be attributed an undue credibility deficit. They may therefore result in apparent testimonial justice, even if injustice may occur through the distortion of conceptual resources in ways that support an inequitable status quo. The specific power dynamics at play when the metaphor is used, as well as the metaphor's relationship to the dominant conceptual resources, must therefore be considered.

For example, consider the “doctor is a military general” metaphor described above. There, I argued that this metaphor, by including a concept in “the physician's role” of someone who holds absolute authority, may be unethical because it reinforces a credibility excess attributed to the physician. Insofar as the conceptual category “physician,” which is used as a marker of credibility, becomes distorted, it increases the credibility of the physician in ways that support the status quo. In general, it may be accurate that physicians enjoy a privileged social position with significant credibility (Parsons 1939) and that this metaphor bolsters this credibility by drawing on the paternalistic conceptualization of the military general (Nie et al. 2016). This metaphor, in general, may therefore sustain false beliefs in excess physician credibility. (For example, a physician may not possess the knowledge needed to be credible on a condition because it is underresearched and undertaught in medical school, and those with lived experience of the condition may have more knowledge about it than physicians, as has been shown to be the case with Chronic Fatigue Syndrome [Blease, Carel, and Geraghty 2017]). When used in the social context in which this is the case, this metaphor may unduly bolster physician testimony, granting them more credibility than is warranted. As the physician enjoys greater privilege in a medical encounter than a patient (Parsons 1939; Blease, Carel, and Geraghty 2017), this metaphor may sustain the status quo of medical paternalism by unfairly distributing epistemic resources in a way that is ethically bad epistemic practice. This injustice can in turn lead to harm through misdiagnosis, and failure to provide appropriate, timely treatment (Manne 2021). This analysis includes only the power relationship between the physician and patient roles.

However, this same metaphor used in the case of a physician experiencing an undue credibility deficit because of their membership in a marginalized group, such as a physician who is racialized, a woman, disabled, LGBTQ+, Indigenous, working-class, and/or an ethnic minority (Kaye 2021; Sibbald 2021; Sibbald and Beagan 2022), may neutralize an existing unjust credibility deficit. Therefore, although the “physician is a military general” metaphor may be ethically bad epistemic practice because it bolsters a physician's testimony to the point of credibility excess, thereby reinforcing a paternalistic ideology, it may also neutralize an undue credibility deficit by challenging racist, ableist, sexist, heteronormative, colonial, and classist ideologies by bolstering testimony to give someone the credibility they are due.

Relatedly, the “born in the wrong body” metaphor may function similarly for those seeking gender-affirming healthcare treatment when the power dynamic between the healthcare provider and person seeking healthcare is biased toward the healthcare provider. This metaphor is the socially dominant metaphor for the transgender experience (Putzi 2017), and one that may need to be employed to achieve credibility in clinical encounters and facilitate gender-affirming medical treatment because of its alignment with the dominant medical discourse (Johnson 2015). When a person is already at a power disadvantage because of their status as patient, using dominant narratives to convey experience may allow for the person to be considered credible (Carel and Kidd 2014). However, on a structural level, it reinforces the idea that gender is binary (Putzi 2017; Bettcher 2014). Not only may this inaccurately reflect the range of experiences of being transgender (Kobabe 2019), it validates and engrains the hermeneutical resources that underscore gender-oppressive power structures within the dominant discourse. When considering the multiple relationships between the physician and the patient within the dominant cis-heteronormative ideology, the metaphor may be ethically bad epistemic practice as it could be considered testimonial smothering, a coerced silencing involving “the truncating of one's own testimony in order to ensure that the testimony contains only content for which one's audience demonstrates testimonial competence” (Dotson 2011, 244), yet it simultaneously enables the person to receive the credibility they are due and access the resources they require.

Similarly, feeling “down” is a common metaphor for depression (Coll-Florit et al. 2021). If this metaphor is being used by a patient to communicate her experience to a health professional in a way that bridges the epistemic gap between patient and healthcare provider, it may facilitate testimonial justice. The patient may be drawing on the hermeneutical resources at her disposal to communicate across an epistemic divide what a healthcare provider may conceptualize as “negative affect.” This may be the best epistemic tool at her disposal and capture her understanding of her experience, resulting in her being received as credible. This is characteristic of testimonial justice (Fricker 2007).

This may be the best epistemic tool at her disposal because a history of Western, colonial medicine, popular culture, and pharmaceutical promotion has entrenched the metaphor “depression is down” and “down is bad” in dominant discourse to the exclusion of others (Elliott 2003; Delbaere 2013; Linklater 2014). It may also limit other possible expressions and conceptually neglect important experiential components, such as loss of appetite, which are in her interest to have attended to. In addition, as feminist critiques of psychiatry have argued, it may circumvent attention to situational components that may justifiably contribute to experiencing “downness,” such as gendered poverty, maternal-role expectations, sexual objectification, and trauma (Tobia et al. 2013). By reinforcing this metaphor through its use in this context, it may entrench a conceptualization of depression that makes those who express alternatives unintelligible and not credible, and therefore be considered an ethically bad epistemic practice at the same time as it facilitates testimonial justice by rendering the testifier credible, which is ethically good practice.

These examples illustrate the complex interaction between testimonial and hermeneutical injustice. In all three of these examples, metaphors were used to ensure the testifier received the credit they were due in situations where an unfair credibility deficit was otherwise present or highly likely to be present. They therefore to appear to be examples of transactional testimonial justice (Anderson 2012); the testifier is drawing on the dominant hermeneutical resources available to capture their experience and is given the reception they are due in the testimonial exchange because this testimony bears the appropriate markers of credibility (Carel and Kidd 2014).

Yet in each of the three cases, the testimony that was provided using metaphor may not have been reflective of the experience the testifier was attempting to communicate. In the example of the physician identifying with a marginalized group, the metaphor may be challenging testimonial quieting, the failure to identify a person as a knower due to their lack of credibility resources (Dotson 2011), by increasing credibility resources at the expense of testimony reflecting experience. In the example of the person seeking gender-affirming care, this may be testimonial smothering: “the truncating of one's own testimony in order to ensure that the testimony contains only content for which one's audience demonstrates testimonial competence” (Dotson 2011, 244). In both cases, although credibility was bolstered, it was bolstered by truncated testimony: testimony only partially reflective of experience. This truncated testimonial justice, whereby a person is given the credit they are due for their testimony, when this testimony insufficiently reflects the experience they are trying to understand and communicate, may be particularly salient with the use of metaphors. Given that metaphors may distort conceptual resources by diluting and pruning resources available within the dominant discourse and that they tend to reflect and reproduce a dominant discourse, their use by those already in a position of social and epistemic disadvantage may increase the likelihood of experiencing truncated testimonial justice. Yet one may not have a choice but to pursue this truncated testimonial injustice along with the material resources it affords as the already inequitable epistemic system would otherwise put them at risk of more severe forms of epistemic injustice. Truncated testimonial justice, although not epistemic justice, may be better than testimonial injustice; it may be a means of survival and a strategy for avoiding further epistemic exploitation in some circumstances (Berenstain 2016).

In the third example, the person with depression gave testimony using metaphor that was not necessarily inaccurate but may have been insufficient to capture the experience of depression. This is likely a case of contributory injustice because the resources that would have been useful have been excluded from the dominant discourse because of the exclusion of their originators (Dotson 2011). The person may have been doing their best with the resources available to them, but these resources are insufficient. Testimonial justice is therefore truncated by the availability of resources with which to testify and maintain credibility.

These three examples illustrate transactional testimonial justice (which I argue is truncated transactional testimonial justice) with simultaneous structural hermeneutical injustice (Anderson 2012). Apparent testimonial justice is achieved in the transaction as the testifier receives the credibility they are due and the material effects this affords, yet the testimony used to achieve this state is not reflective of their experience because of the inequitable distribution of epistemic resources within the epistemic system. Their testimony may continue to entrench the credibility of dominant discourses and the lack of credibility of alternatives, and yet may be more ethical than the injustice that would otherwise result. It may be the better of two unjust options. Examples like these require a case-by-case assessment with close attention paid to all effects, particularly if epistemic justice in its fully realized form is unavailable because of the inequitable epistemic system within which testifiers find themselves.

In addition, the inverse is also possible: hermeneutical justice may be achieved at the expense of transactional testimonial justice. Although those who use novel metaphors may help to create and sustain the hermeneutical resources needed to challenge oppressive systems, the cost is the potential for transactional testimonial injustice—these metaphors may render one not credible. K. Steslow describes this type of experience when discussing her use of the metaphor “experiencing mental illness is being a migratory bird experiencing avian flu” (Steslow 2010). This metaphorical description is at odds with how those with whom she interacts in the medical system conceptualize and understand mental illness—it does not have the markers of credibility in the medical context because it challenges the dominant discourse (Carel and Kidd 2014). By using this metaphor, she risks that it will contribute to the perception that she is not credible, which is reinforced by the stereotypes of those experiencing mental illness (Crichton, Carel, and Kidd 2017). The result for Steslow is very real, material harm, such as continued hospitalization and dehumanization, as this lack of credibility serves as a marker of lack of rehabilitative progress (Steslow 2010). Although her description may be creating a hermeneutical resource that reflects her experience of mental illness and thereby facilitate hermeneutical justice, its use renders her not credible, creating transactional testimonial injustice and exposing her to material threat.

I have argued thus far that metaphors may have epistemic consequences for both the terms they employ and the concepts they wish to illuminate. These consequences result from possible dilution, where concepts are expanded to contain ideas that they otherwise would not, and exclusion, where elements of concepts are pruned or “pushed out” of a conceptual space. The result is the potential for metaphors to lead to hermeneutical injustice if they shape concepts in the dominant discourse, either by dilution or exclusion, so that they are no longer sufficient for those who have been epistemically marginalized to convey their experience. However, metaphors also have the potential to facilitate hermeneutical justice by diluting problematically reduced concepts in the dominant discourse to make them more useful for those who have been epistemically marginalized to convey their experience. In addition, because metaphors can shape concepts of credibility, metaphors also have the potential to lead to testimonial injustice. If metaphors distort concepts in the dominant discourse in ways that make it more difficult for accurate credibility judgments to occur, they may lead to testimonial injustice. They may also, however, distort already problematic concepts in the dominant discourse to facilitate accurate credibility judgments. This testimonial in/justice is the result of the hermeneutic effects of metaphors on a conceptual level. In addition, testimonial injustice may occur if the act of distorting the hermeneutical resources available through metaphor is perceived as a marker of lack of credibility. This perception may lead to testimonial injustice regardless of the hermeneutical benefits or harms on a conceptual level. Both the social and conceptual work metaphors do have ethical and epistemic consequences.

VI. Making Judgments—Is This Metaphor Ethically Bad Epistemic Practice?

Based on this analysis, clearly many factors are involved in considering whether a metaphor is ethically bad epistemic practice. To assess these factors, I propose asking these questions:

1. 1. Is the person using the metaphor, who exists within an inequitable epistemic system, using the best epistemic resources at their disposal to communicate their experience?

   If yes, then it is unlikely that the metaphor is ethically bad epistemic practice.

   1. 1.1 If the person is not using the best resources at their disposal to communicate their experience, would using the “best” resources increase their likelihood of experiencing epistemic or material harm?

      If yes, then it is unlikely the metaphor is ethically bad epistemic practice.

2. 2. Is the metaphor altering a concept that acts as a marker of credibility within the social context in which it is being used?

   If yes, the metaphor has the potential to be ethically bad epistemic practice.

   1. 2.1 Is the metaphor altering a concept that acts as a marker of credibility in a way that results in an increased likelihood of someone experiencing a credibility error?

      If yes, it is likely the metaphor is ethically bad epistemic practice.

   2. 2.2 Is the metaphor altering a concept that acts as a marker of credibility in a way that moves toward correcting an existing credibility error?

      If yes, it is unlikely the metaphor is ethically bad epistemic practice.

1. 3. Is the metaphor altering one of the concepts it uses when this concept is used to capture a group's lived experience within the dominant discourse?

   If yes, it is likely that the metaphor has epistemic and ethical consequences.

   1. 3.1 Is the metaphor diluting a concept in the dominant hermeneutical resources so that it detracts attention from key components that capture the experience for some or all members of an epistemically marginalized group?

      If yes, it is likely the metaphor is ethically bad epistemic practice.

   2. 3.2 Is the metaphor excluding from a concept in the dominant hermeneutical resources components that capture the experience of some or all members of an epistemically marginalized group?

      If yes, it is likely the metaphor is ethically bad epistemic practice.

2. 4. How likely is it that this metaphor will be taken up within the dominant discourse, or to reflect the ideologies of the dominant discourse?

   Metaphors likely to be taken up by, and that support, the ideologies of the dominant discourse in an inequitable epistemic system likely constitute ethically bad epistemic practice. Metaphors that likely will be taken up by the dominant discourse, but challenge its ideologies, are unlikely to constitute ethically bad epistemic practice.

The first question draws attention to two elements of metaphors in context. First, it accounts for the inequitable epistemic system in which conveyers of information find themselves. Insofar as this system has an ongoing history of perpetuating structural hermeneutical injustice, individuals are not at fault for the absence of epistemic resources available to convey their experience within the dominant discourse. In the example above, Steslow is not at fault for using a metaphor (mental illness as avian flu) that is not understood when the dominant discourse, because of an ongoing history of sanism, has systematically excluded the resources needed to do so, characteristic of contributory injustice (Dotson 2012). Similar to “white ignorance” (Mills 2007), the absence of adequate knowledge of a group's experience from the dominant discourse is not the fault of those who fail to be understood within an inequitable system. Similarly, it is not necessarily the fault of a receiver for failing to understand the metaphor when the reason for this is structural hermeneutical injustice (Anderson 2012). Structural hermeneutical injustice may be operating in the background, and become obvious during the use of the metaphor, but it is not the metaphor itself but the social epistemic structures that are causing the injustice.

Second, this criterion highlights a distinction between using a metaphor to convey one's personal experience and using a metaphor in a way that shapes the concept of a group's experience in general. These are not necessarily mutually exclusive because repeated use of a metaphor will shape the concept as it is used by many people (Lakoff and Johnson 2003). However, the possibility remains that the best resources at one's disposal to convey their experience may be problematic for any or all of the other reasons listed, and calling the use of these resources for this purpose unjust is inherently counter to epistemic justice. In addition, this is disproportionately likely to affect those who are epistemically marginalized, who are less likely to have effective resources in the dominant discourse to convey their experience (Mills 2007). Expression of one's experience through the best resources at one's disposal cannot be unjust. Injustice is likely to emerge, however, when one is using metaphors to convey an experience that is not theirs, or that generalizes the experience of a group with diverse experiences. This occurs because metaphors can shape concepts in ways that may then deny others the hermeneutical resources they need to convey their own experience.

Question 1 continues by considering the harm that can arise from using the best epistemic resources available within an unjust epistemic system. This may be particularly salient when one is in a position of marginalization such that the best epistemic resources one has at their disposal are counter to, or alter, the dominant conceptual resources in contexts where this is unsafe. Not accounting for this circumstance opens those in positions of marginality up to epistemic exploitation by compelling them to educate members of the dominant group on the experience of marginality (Berenstain 2016). Protecting oneself from exploitation is not unethical. In addition, this question allows for consideration of the credibility errors that can result from challenging and altering the dominant discourse. As in Steslow's case described above, using the best epistemic resources at her disposal to describe her experience, that of a bird with avian flu, in the context of psychiatric hospitalization, increased the likelihood that she would be judged not credible for speaking outside the dominant recovery metaphor. This not only increased the likelihood of not only the epistemic harm of experiencing testimonial injustice but also increased the likelihood of a variety of material harms. Choosing to use the dominant metaphor in this context to decrease the likelihood of experiencing a credibility error and the injustices and harms that result would not be unethical. Within epistemic systems that perpetuate harm as their status quo, protecting oneself from this harm and experiences of epistemic injustice is not unethical. Even if a metaphor is likely to be unethical for other reasons, when it is used because it is the best epistemic resource available or to protect oneself from injustice resulting from an inequitable epistemic system, it should not be considered unethical.

Importantly, this question does not necessarily provide criteria by which we can judge the ethics of others’ metaphors. Particularly with question 1, without significant additional information, one cannot know if another person is using the best epistemic resources at their disposal to convey their experience. However, one can reflect on whether they themselves are using the best epistemic resources at their disposal, or uncritically repeating metaphors of the dominant discourse as their experience.

Question 2 interrogates the metaphor's effect on markers of credibility, distinguishing between metaphors that shape concepts to create or exacerbate existing credibility errors in the dominant discourse, and those that correct for errors that exist because of structural group-based credibility errors. Given that credibility errors are part of the dominant discourse, shaping this discourse to correct for these errors promotes epistemic justice. Contrarily, perpetuating the miscognition of credibility errors embedded in the dominant discourse, such as by comparing those with dementia to zombies, perpetuates the status quo in a way that is likely to constitute ethically bad epistemic practice.

Question 3 interrogates metaphors’ effects on the concepts for capturing experience, differentiating between those that shape the dominant hermeneutical resources in ways that are more or less useful for supporting those who have been epistemically marginalized to convey their experience. Making concepts less reflective of experience and less useful facilitates miscognition in ways that support an inequitable status quo, as occurs with rape metaphors, thereby constituting ethically bad epistemic practice.

Question 4 interrogates the metaphor in relation to the dominant discourse and ideologies that support it. It does not act, in itself, as a qualifier on which to judge a metaphor, but rather, in conjunction with what is illuminated by interrogating the metaphor using questions 1 through 3, enables one to interrogate the likely magnitude of possible harm. Those in positions of power or privilege are more likely to have the metaphors they use taken up by the dominant discourse (Lakoff and Johnson 2003). For example, a white, heterosexual, cisgender, upper-class, male political leader who uses a metaphor that is then repeated by numerous media sources is likely to have more impact on shaping the conceptual resources in the dominant discourse than a person using the same metaphor in a casual conversation with one other person. The likelihood of this conceptual alteration becoming dominant discourse because of the user's social and epistemic positionality is an important consideration. Those in positions of power and privilege therefore have more responsibility to interrogate the epistemic effects of the metaphors they use because the likely epistemic impact resulting from their influence on the dominant discourse is greater. As well, metaphors that support dominant ideologies are also more likely to become dominant within a discourse (Efeoğlu Özcan 2022). They are therefore more likely to shape conceptual categories in ways that may promote epistemic injustice. Considering what ideology is facilitated by the alterations in concepts metaphors promote—such as the ageist and ableist ideology of the “people with dementia are zombies” metaphor, and the misogynist ideology of the “men are wolves” metaphor—is an important ethical consideration. The risk of harm is greater when the metaphor supports an ideology that underlies the dominant discourse because of the inequitable epistemic system in which it is grounded. Metaphors that challenge these ideologies are likely those working toward epistemic justice.

VII. Unifying the Conversation—Illness as War

In these cases, a definitive conclusion as to whether a metaphor constitutes ethically bad epistemic practice is complicated, but these questions guide reasoning about the ethical and epistemic implications of metaphors and the various ways epistemic injustice is reproduced and challenged by them. This process helps to unify ongoing discussion about the ethics of a highly debated metaphor: the “illness is war” metaphor that is commonly employed in healthcare (George, Whitehouse, and Whitehouse 2016; Nie et al. 2016; Tate and Pearlman 2016; Sontag 1990a; 1990b; Chapman and Miller 2020; Tate 2020).

The tension arising between transactional testimonial injustice and structural hermeneutical injustice is the tension that, in different words, underscores the primary ethical debate around metaphor use in healthcare. On the one side are those arguing that war metaphors reinforce oppressive systems at odds with the values medicine seeks to promote and that such metaphors also limit the possibilities for understanding, communicating about, and responding to, illness (Sontag 1990b; George, Whitehouse, and Whitehouse 2016; Nie et al. 2016). This is effectively an argument that metaphors promote hermeneutical injustice. On the other are those who argue metaphors can be empowering and facilitate communication, particularly when the patient and healthcare provider do not have a shared language (Tate and Pearlman 2016; Tate 2020). This is effectively an argument that they promote testimonial justice. Although the questions outlined here do not resolve this debate, they bring to light the importance of considering the inequitable epistemic system and the intersectional power relationships that shape the context in which the metaphors are used. Both sides are correct, but both fail to acknowledge the work this metaphor does within a broader, epistemically unjust system, and how this work changes based on those using this metaphor and their relationship to the dominant discourse. I therefore aim to shift the conversation from whether the metaphor is inherently ethically bad epistemic practice to how one may negotiate epistemic justice given the metaphor's potential to promote both epistemic injustice and justice in different contexts. Given that metaphors do epistemic work, how can we use them to do ethical epistemic work?

VIII. Negotiating Boundaries

An important caution, though, is that even if a metaphor may be considered ethically bad epistemic practice for reasons outlined above, this does not mean the use of this metaphor should necessarily be stopped, or the metaphor abandoned. Based on the analysis provided here, it is possible for metaphors that constitute “ethically bad epistemic practice” because they facilitate miscognition, supress nondominant discourse, and reinforce the status quo, to also help to challenge these same things. If any metaphor that has unjust consequences is abandoned, regardless of whatever other justice it may facilitate, this may result in disproportionate harm to those in already epistemically disadvantaged groups, and itself be considered ethically bad epistemic practice. An argument to abandon metaphors because they result in epistemic injustice could be used to justify constraining language that has the potential to be used for resistance because it supports the status quo in some ways, although it challenges it in others, depending on the context. It could be used to silence voices who are attempting to communicate their experience in the only way they know how to or the only way in which others will listen. It has the potential to erase the consideration of intersectional experiences and multidirectional power relationships that is in the interest of (at least) disadvantaged groups to have understood. Caution must be taken to ensure this is not the case.

As such, while facilitating miscognition, distorting concepts, and reinforcing the status quo may be considerations in determining whether a metaphor is ethically bad epistemic practice, through an intersectional lens, they may not be sufficient justification for abandoning a metaphor. Given the multiple ways cognition and epistemic injustice are related, as well as the multidirectional power relationships involved in intersectional experience, if we do not have a frame of reference in which these can be considered during reasoning, we risk committing the same injustice we seek to identify: we risk perpetuating ethically bad epistemic practice. I argue that this analysis provides epistemic resources to reason through these debates in ways that foreground the experience of those epistemically marginalized.

IX. Metaphors as Epistemic Resources within Inequitable Epistemic Systems

As Gaile Pohlhaus Jr. describes: “Good epistemic resources put us in particular relation to our experiences (for example, noticing more or certain kinds of details about the experience or anticipating what will follow from the experience). If our language, concepts, or standards don't do that, then we need to develop new resources that do” (Pohlhaus Jr. 2012, 718). Metaphors have the potential to be good epistemic resources. They direct us to notice more or certain kinds of details, but at the same time, they obscure others. Within an unjust epistemic system, metaphors that support the dominant discourse and subsequently the ideologies that underly it are also those that are likely to shape the concepts they use in ways that make them less useful for those in marginalized groups to convey their experience. This is because, as Lakoff and Johnson describe, “The canonical person forms a conceptual reference point, and an enormous number of concepts in our conceptual system are oriented with respect to whether or not they are similar to the properties of the prototypical person” (Lakoff and Johnson 2003, 132). Those who deviate from what the dominant discourse considers “prototypical” are already epistemically disadvantaged because the hermeneutical resources available through these concepts are less likely to capture their experience. Altering the resources that do, so they are no longer useful for this purpose, constitutes ethically bad epistemic practice by promoting miscognition that supports and exacerbates an inequitable status quo.

Within this inequitable epistemic system, metaphors that perpetuate the status quo by altering concepts that can be used to convey the experience of those marginalized by this system may be unethical. Metaphors can do this by excluding from the conceptual frame of reference components of experience relevant to those who are epistemically marginalized by the dominant discourse. They can also do this by diluting concepts so that they no longer make salient important components of experience that are in someone's interest to convey. Both the concepts used and those produced by a metaphor are vulnerable to these alterations. In addition, because concepts are used as markers of credibility, metaphors can shape concepts to perpetuate credibility errors by excluding or diluting conceptual components.

These effects highlight two important theoretical considerations for epistemic injustice. First, hermeneutical injustice does not necessarily constitute a gap in hermeneutical resources. It can also occur when existing hermeneutical resources are altered through dilution and exclusion so that they no longer become useful concepts for conveying experience. In these circumstances, rather than leaving a gap in resources, resources become pruned or washed out of the dominant discourse. It is, in effect, the removal of available epistemic resources to convey experience, which can be just as harmful. Second, because the removal of resources is highlighted, I argue it is not only important to add conceptual resources to the dominant discourse relevant to capture the experience of those who are experiencing ongoing epistemic injustice, but also preserve those that are already there and useful and at risk of distortion in ways that may constitute ethically bad epistemic practice.

Although metaphors have the potential to cause epistemic harm, they also have the potential to work toward epistemic justice by altering existing concepts in ways that better convey experience, particularly the experiences of those who are epistemically marginalized. A metaphor's relationship to the dominant discourse, to the experience of the person conveying it, as well as how it shapes conceptual resources are all important considerations in determining when a particular metaphor is unethical. Importantly, because of structural hermeneutical injustice inherent in the epistemic system, in defining a metaphor as inherently unethical, we risk blaming or silencing those already disadvantaged by the epistemic system within which they survive. Structural hermeneutical injustice, and the experiences it promotes, such as testimonial injustice, hermeneutical injustice, epistemic exploitation, and ignorance, must therefore be central considerations in the discussion of the ethics of metaphors to decrease the likelihood of these harms.

By influencing our conceptual systems, metaphors not only shape our epistemic resources, but influence how we act in the world and the reality against which we measure “truth” (Lakoff and Johnson 2003). When metaphors distort concepts in dominant discourses in ways that promote epistemic injustice, they risk not only perpetrating epistemic and material harm, but legitimating this harm within the framework for reasoning they evoke. Because our conceptual system is both inequitable and in constant flux, interrogating how we engage with and shape this system, and the consequences, is an ethical task.