1. INTRODUCTION
A first step to fighting injustice is coming to understand our social experiences well enough to be able to make them intelligible to others. It is not unusual for victims of injustice to describe a stage at which they have some sense of being wronged, but lack the language – and the concepts – to articulate that wrong to others. In many such cases, there is something missing from our shared linguistic/conceptual resources that makes it difficult to capture what is happening. This gap arguably affects all who share those resources. But the lion's share of the disadvantage falls upon those suffering the wrong, insofar as they are prevented from speaking out about it, and from politically organizing to address the injustice.
In this paper, I'll focus on a series of cases that have the features described above. I'll use these cases to raise three closely related questions. The first question is about the lack in our shared linguistic/conceptual resources. What do the gaps in these resources consist in? Or to put it another way, what precisely is missing in these cases? The second question is about what it takes to overcome the gaps in our shared resources. What does it take to fill in these gaps? And the third question is about the nature of the achievement involved in overcoming these gaps. What kind of achievement is this, and specifically, is it a cognitive achievement?
In response, I'll defend two main claims. First, I'll argue that the gaps in my central cases are filled in, at least initially, by labels of a particular kind. Specifically, these are labels that allow us to see the experiences in question as novel instances of phenomena with which we are already normatively familiar, in a sense to be explained below. But second, I'll also argue that these labels bring with them important downsides: they introduce distortions into our understandings of the experiences. Using a pair of case studies (‘statutory rape,’ ‘sexual harassment’), I'll illustrate these distortions. The fact of these distortions at least complicates any effort to argue that these cases represent cognitive achievements in a straightforward sense. Further, recognizing these distortions helps us appreciate a further harm of certain injustices, namely, that they put their victims in the position of having to choose between communicating their experiences to others, via generally available labels, and genuinely understanding those experiences.
Some of my central cases are drawn from Miranda Fricker's influential work on hermeneutical injustice (Fricker Reference Fricker2007, especially Ch. 7). I follow Fricker in thinking that these cases genuinely involve gaps, or lacunae, in our shared linguistic/conceptual resources. Additionally, I consider Fricker's suggestion that filling in a gap of this sort constitutes a “cognitive achievement,” and often a significant one at that (Fricker Reference Fricker2007: 148). But it is not my aim in this paper to offer an account of hermeneutical injustice as such. As will become clear shortly (Sections 2–3), the literature on hermeneutical injustice includes cases that are importantly different from my central ones, and I take no position on whether the label ‘hermeneutical injustice’ should be used broadly (to include all or most of the cases in the literature), or more narrowly.
This paper is part of two larger projects. One project seeks to better understand when our language and our concepts are adequate for our normative purposes. More carefully, since much of our linguistic/conceptual repertoire is adequate in some respects and less so in others, my aim is to distinguish some different dimensions of evaluation for that repertoire. I am particularly interested in the extent to which our linguistic/conceptual resources succeed (or fail) to normatively illuminate our social experiences, and how failure in this respect can differentially burden the most vulnerable in our society, thereby reinforcing their place in our social hierarchy.
A second, related, project is to think further about what it takes to overcome epistemic, linguistic, and related injustices. In recent years, significant progress has been made in identifying and better understanding injustices of all of these types. I'd like to focus on what is involved in ameliorating them as well. Perhaps unsurprisingly, the picture that will emerge is that the process of amelioration is a long and arduous one, where genuine achievements are regularly accompanied by significant setbacks. I'll offer some tentative suggestions about why that might be in later sections of this paper.
The paper proceeds as follows. In Section 2, I describe my three central cases. I also briefly present Fricker's view of hermeneutical injustice, and emphasize how broad her characterization of the injustice is. In Section 3, I consider three proposals about the gaps involved in my central cases, and defend one of them. I also consider some pros and cons of my preferred proposal. Then, in Section 4, I turn to some case studies in order to describe the distortions introduced by efforts to overcome gaps. I conclude in Section 5 by taking stock of the discussion in this paper, and raising some questions for future consideration.
2. THREE CENTRAL CASES
Fricker introduces her notion of hermeneutical injustice via an account of a consciousness-raising session, drawn from Susan Brownmiller's history of the feminist movement of the 1960s and 70s (Brownmiller Reference Brownmiller1999: 180–5). Brownmiller tells the story of a woman, Wendy Sanford, who had been fighting depression following the birth of her son. A friend drags Sanford to a session on women's medical and sexual health at a local university. The following is Sanford's own recollection of one part of that experience:
In my group people started talking about postpartum depression. In that one forty-five-minute period I realized that what I'd been blaming myself for, and what my husband had blamed me for, wasn't my personal deficiency. It was a combination of physiological things and a real societal thing, isolation. (Brownmiller Reference Brownmiller1999: 182, quoted in Fricker Reference Fricker2007: 149)
Sanford's isolation, combined with other factors like the general failure to talk about the ubiquity of depression, had prevented her from making sense of what was happening to her. According to Fricker, Sanford thereby suffered a hermeneutical injustice.
For Fricker, a hermeneutical injustice is “the injustice of having some significant area of one's social experience obscured from collective understanding owing to hermeneutical marginalization” (Fricker Reference Fricker2007: 158). That obscuring is sometimes due to gaps in our shared linguistic/conceptual resources. This picture fits well with Sanford's case: she (Sanford) might not have suffered the harms she did had ‘postpartum depression’ been available to her as a way of capturing what was happening to her. But note that the definition quoted above doesn't require hermeneutical injustices to always involve such gaps. That is, as far as that definition is concerned, other things beside linguistic/conceptual lacunae can also qualify as sources of hermeneutical injustice when they interfere with collective understanding. In this respect, the definition is strikingly broad.
Fricker argues that a primary, or “essential,” harm of hermeneutical injustice is a particular kind of communicative disablement (Fricker Reference Fricker2007: 44, 162). The agent who suffers a hermeneutical injustice is prevented from rendering her experience “communicatively intelligible” to relevant others, even though she has a strong interest in doing so. Where the injustice is due to a gap in our linguistic/conceptual resources, that gap also interferes with the agent's communicative goals.
This notion of communicative intelligibility will be important in my later discussion, so let me say a little more about it here. (Fricker doesn't unpack this notion much further in her discussion.) Even before attending the session on women's health, Sanford was presumably able to describe some of what was happening to her. Though Brownmiller doesn't offer details, we can imagine that Sanford could explain that she felt constantly tired, disaffected, unable to contribute to her son's care, and so on. What she couldn't do was describe her experience in a way that made its crucial normative properties clear to her interlocutors. In particular, she couldn't describe her experience in a way that would help explain why she wasn't solely to blame for her behavior. That suggests that to be prevented from making an experience communicatively intelligible in the relevant sense is to be prevented from describing that experience in some way that grounds its significant normative properties. Which normative properties those are will vary from case to case.Footnote 1
A second case of hermeneutical injustice – and the case that is most often discussed in the literature on this topic – also comes from Brownmiller's work. It involves what we would now label ‘sexual harassment,’ before that phrase was in widespread use. The case involves a woman, Carmita Wood, who was a staff member at Cornell University (Brownmiller Reference Brownmiller1999: 279–85). Wood was subjected to severe mistreatment by a senior professor in her department, over a significant period of time. The professor made lewd advances towards Wood, groped her, cornered and forced sexual contact on her, and more. As a result, she developed a range of physical symptoms, including chronic back and neck pain; eventually, she was forced to quit her position.
Wood later connected up with some local lawyers and activists interested in helping her, and others with similar experiences, speak out. But speaking out required describing what had happened, and at that point, that was a major stumbling block. Brownmiller quotes one of those involved in the speaking-out effort in the passage below:
The ‘this’ they were going to break the silence about had no name. “Eight of us were sitting in an office of Human Affairs,” Sauvigne remembers, “brainstorming about what we were going to write on the posters for our speak-out. We were referring to it as “sexual intimidation,” “sexual coercion,”, “sexual exploitation on the job.” None of these names seemed quite right. We wanted something that embraced a whole range of subtle and unsubtle persistent behaviors. Someone came up with “harassment.” Sexual harassment! Instantly we agreed. That's what it was.” (Brownmiller Reference Brownmiller1999: 281, quoted in Fricker Reference Fricker2007: 150)
The efforts of Wood and her colleagues, as well as other feminist activists and lawyers, led to significant legal reforms in the U.S. Perhaps most notably, these included the issuance in 1980 of guidelines by the Equal Employment Opportunity Commission defining sexual harassment in the workplace, and declaring it to be a form of sex discrimination. I'll return to these legal reforms in Section 4.
Several features of Brownmiller's account are noteworthy. First, those brainstorming in the scene she describes consider several related labels – ‘sexual intimidation,’ ‘sexual coercion,’ etc. – for Wood's experience, but reject most of them as inadequate. Second, when they do alight upon their preferred label – ‘sexual harassment’ – there is something that might be described as a ‘Eureka!’ moment. Presumably, something about this label seems to fit what happened to Wood in a way that the other labels do not, and this is apparent to all concerned.Footnote 2 Finally, that last feature is arguably present in the Sanford case as well: Sanford's reaction when she first hears the phrase ‘postpartum depression’ can also be described as a ‘Eureka!’ moment.
There is something at least prima facie surprising about these ‘Eureka!’ moments. Their occurrence suggests that something significant is achieved in the conversations recounted in the Sanford and Wood cases. Recall that, qua instances of hermeneutical injustice, these are meant to be cases in which understanding is curtailed, perhaps due to gaps in our linguistic/conceptual resources. But achieving understanding typically takes time and effort. That is, it's not typically something that happens rapidly, such as over the course of a single conversation. That raises the following question. What is accomplished in the conversations featured in the two cases, that gives rise to these striking ‘Eureka!’ moments?Footnote 3 I'll return to this question in Section 3.
I'll turn now to the third of my central cases, having to do with statutory rape. Very roughly, statutory rape is sexual intercourse with someone under the age of sexual consent. The label ‘statutory rape’ is of relatively recent vintage, coming into widespread usage only around 1900.Footnote 4 (Compare this with ‘rape’, which has been in general use for a much longer time.Footnote 5) As in the case of sexual harassment, the advent of the label ‘statutory rape’ coincided with significant legal reforms, which I'll briefly sketch below.
From the colonial period onward, statutory rape laws in the U.S. were gender-specific, prohibiting sex with girls, but not boys, under the age of consent (Cocca Reference Cocca2004, especially Ch. 1). The age of (sexual) consent during this time generally varied between 10 and 12 from state to state. In the late 19th century, just around the time that ‘statutory rape’ was coming into broad usage, a coalition of feminist activists, religious leaders, and working-class men's organizations came together to lobby to raise that age of consent. Their efforts were successful: in nearly every state, the age was raised to between 16 and 18. Other changes happened around the same time, partly in response to this increase. For example, many states began to codify into law that previous sexual experience on the part of the (female) victim could serve as a defense against a charge of statutory rape.
It's plausible to think that the increased availability of ‘statutory rape’ in shared linguistic/conceptual resources played some role in overcoming a hermeneutical injustice, specifically by garnering recognition that sexual intercourse between adults and young teenagers can constitute a serious wrong. As mentioned above, intercourse between adults and younger children had already been illegal in much of the U.S. prior to the legal reforms described above; these reforms had the effect of extending the prohibition to intercourse with older children. Understanding intercourse between, say, a middle-aged man and a 14-year-old girl as a kind of rape can help provide a strong rationale for prohibiting that conduct.Footnote 6
At the same time, it's important not to be too Pollyannaish about either the motivation, or the impact, of the changes described above. The reformers seeking to raise the age of consent were driven by several distinct motivations, some of which had significant racist or classist overtones. Carolyn Cocca argues that while some reformers sought to undo a sexual double standard, where women but not men were penalized for sexual activity prior to marriage, others were driven by a desire to control sexual behavior by working-class women in particular. She also notes that some black women's groups strongly opposed the reforms, fearing both that they would not help protect black women, and that they would be used to further target black men (Cocca Reference Cocca2004: 13–14). I'll return to these and later reforms in statutory rape law in Section 4.
We now have all three of the cases that will be central in the remainder of my discussion. The literature on epistemic injustice contains several further phenomena that have been argued to also be instances of hermeneutical injustice. To mention two particularly noteworthy examples: José Medina has argued that white ignorance in the sense characterized by Charles Mills fits Fricker's definition of hermeneutical injustice (Medina Reference Medina2013, especially Ch. 3; see also Mills Reference Mills, Sullivan and Tuana2007 for his account of white ignorance); and Katherine Jenkins has argued that some rape myths and some domestic violence myths also fall in the same category (Jenkins Reference Jenkins2017). I'll briefly touch on these further cases in the next section.
3. WHAT'S MISSING?
I turn next to three proposals regarding the nature of the gap in our linguistic/conceptual resources. According to the first proposal, the gap consists in a missing label (of any kind) for the experience in question. According to the second proposal, the gap consists in widely prevalent false beliefs that interfere with a full understanding of the experience. And according to the third proposal, the gap again consists in a missing label, but now a label of a particular kind, one that renders the experience communicatively intelligible.
As I'll argue below, the first proposal is a non-starter. But it will be useful to briefly discuss it nonetheless, in order to get a sense of what a better proposal might look like. The second and third proposals each capture a philosophically interesting range of cases; but I'll argue that we have good reason to prefer the third one with respect to the three central cases. This list of options is by no means exhaustive; and further options may well be worth considering. But I won't try to touch on every possible variation here. Rather, my main aim in this section is to highlight the virtues of the third proposal, particularly the fact that that proposal, unlike the other two, can provide a unified account of my central cases.
Let's begin with the first proposal, according to which the gap consists in a missing label, of any kind.Footnote 7 On this proposal, filling in the gap requires making available a label for the experience in question. Some of Fricker's discussion suggests something like this thought, such as when she writes that the Wood case features a “lacuna where the name of a distinctive social experience should be” (Fricker Reference Fricker2007: 150–1). And it's easy to see why this proposal seems like a step in the right direction. As we saw earlier, a primary harm of hermeneutical injustice is a kind of communicative disablement. A label can serve as a vehicle for communication about the experience.
Nevertheless, this first proposal clearly fails, for several related reasons. First, as Laura Beeby points out, there were labels available for Wood's experience (Beeby Reference Beeby2011: 480). ‘Flirting’ and ‘chasing around the desk’ might have been among the most readily accessible. But the availability of these labels did not prevent the hermeneutical injustice. Second, that's because not just any label will help make sense of the experience. If anything, the labels just mentioned were a big part of the problem facing Wood, for application of those labels might suggest that she was overreacting to the professor's behavior. (Indeed, in Brownmiller's telling, Wood was for a long time embarrassed about her own reaction to the professor's behavior.) Finally, not just any label will make the experience communicatively intelligible. Even if any label can enable communication about an experience, not just any label will lay bare its crucial normative properties, as is required for communicative intelligibility.
Let's move now to the second proposal, according to which the gap consists in widespread false beliefs. More specifically, these are beliefs that prevent us from developing a full understanding of the experience in question.Footnote 8
This second proposal fits some cases in the epistemic injustice literature very well. Recall, in particular, Jenkins’ view that rape myths sometimes constitute hermeneutical injustice. These myths include widespread false beliefs to the effect that “rape always involves overwhelming physical force, and that victims of rape always physically resist their attacker,” among others (Jenkins Reference Jenkins2017: 192–3). Jenkins argues that these myths have many pernicious effects. For one, those who accept them are especially unlikely to recognize that they have been raped, if the rape occurs in circumstances that align with the myths they accept. In this and other ways, acceptance of these myths occludes rape from collective understanding. Further, Jenkins also suggests that the prevalence of rape myths means that (many of) our operative concepts of rape are defective ones (Jenkins Reference Jenkins2017: 195–7; see also Haslanger Reference Haslanger2012b for the distinction between operative and manifest concepts).
On the proposal we are currently considering, filling in the gaps in our linguistic/conceptual resources will require dispelling the widely held false beliefs. But it may well require much more than that: for example, it may also require developing non-defective operative concepts that apply to the experiences in question.
This second proposal is clearly more promising than the first one. The prevalence of rape myths (and domestic violence myths) is clearly an injustice, and this proposal helps get at why it should qualify as a hermeneutical one.
Nevertheless, there are difficulties with this proposal as well. To begin with the milder of these: it's very likely that we have false beliefs about most of our social experiences, and that many of these beliefs prevent us from fully understanding those experiences. If that's right, then on the current proposal, hermeneutical injustices will prove to be extremely commonplace.Footnote 9
Perhaps this isn't such a bad result. Even if hermeneutical injustices are everywhere, perhaps we can still say that some are more significant than others, along various axes. So we needn't be equally concerned with all injustices in this category. Still, insofar as there is something distinctive about my central cases – both with respect to the injustice suffered, and with respect to what overcoming them requires – that distinctiveness isn't captured on the current proposal.
Alternatively, it could be argued that not every collection of false beliefs about a social experience qualifies as a hermeneutical injustice, because not all such beliefs render the associated concept as defective. That's true for any concept, not just social experience concepts. To illustrate: even if I share the widespread false belief that ostriches hide their heads in the sand when threatened, that isn't – by itself – enough to make my ostrich concept problematic in any interesting sense. By contrast, sufficiently many false beliefs about ostriches, or false beliefs of the right kind, would indeed render that concept defective. Now recall that on Jenkins’ view, rape myths constitute our operative concepts of rape as defective ones. Even if that's right, we needn't infer that just any false belief(s) about rape would have the same consequence; perhaps it's something about the particular contents of rape myths, or their centrality to our operative concepts, that explains why they render those concepts as defective.
This response seems to me to be on the right track. But note that it only helps prevent over-proliferation of hermeneutical injustices when paired with an account of when false beliefs render an associated concept as defective, and when they don't; and that's at least a challenging task.
A more significant difficulty for the current proposal is that it fails to capture why there are hermeneutical injustices in my central cases in the first place. To see this, let's focus on the postpartum depression (Sanford) case. In Brownmiller's telling, Sanford (and her husband) clearly had some false beliefs about what was happening to her: they thought, for example, that her inability to care for her newborn son was her own “personal deficiency.” But suppose that they had lacked those false beliefs. Even without them, Sanford would still have suffered a hermeneutical injustice, in much the same way as she does in the actual case. That is, she would still have been unable to make sense of what was happening to her. Further, she would still have been unable to render her experience communicatively intelligible to relevant others, such as her husband; and that, in turn, would have meant that she would have continued to not get the help she needed. That's reason to think that the false beliefs are incidental to the hermeneutical injustice in the Sanford case, unlike in Jenkins’ cases.Footnote 10 (Similar points can be made about the sexual harassment (Wood) and statutory rape cases, though I won't rehearse those details here.)
Let me emphasize here that I'm not arguing that rape myths and domestic violence myths should not be regarded as hermeneutical injustices. As mentioned in the Introduction, it is not my purpose in this paper to defend a view of hermeneutical injustice as such. Rather, my point here is only that there are interesting differences between the cases captured by the second proposal and the three cases that are central to this paper. Even if they all qualify as hermeneutical injustices, it's worth noting the differences between them.
We can move now to the third, and final, proposal. According to this proposal, the gap in our linguistic/conceptual resources consists in a missing label, of a particular sort. Specifically, what's missing is a label that renders the experience communicatively intelligible in the sense explained in Section 2. And the most straightforward way to render an experience communicatively intelligible is to assimilate it to another whose relevant normative properties are already sufficiently familiar.
There are several ways in which this kind of assimilation can happen. (i) One way it can happen is by co-opting an already normatively familiar label, and extending it to a further experience (or range of experiences). ‘Stalking’ being used to describe certain forms of online behavior, say, on Facebook or Twitter, is an example of this first type. (ii) Another way is by modifying a label that's already normatively familiar. ‘Postpartum depression’ and ‘statutory rape’ are examples of this second type; each is a modification of a label (‘depression,’ ‘rape’) of older vintage. (iii) Yet another way is by coining a new label, taking as its paradigmatic instance some normatively familiar experience, and applying it to a further experience (or range of experiences). ‘Sexual harassment’ is arguably an example of this third type, insofar as it takes as paradigmatic cases in which sexual favors are demanded in exchange for some benefit (e.g., keeping a job), but also includes further cases in which there is no such quid pro quo, explicit or otherwise.Footnote 11 And there are no doubt further ways to carry out this assimilation as well.
In each of the cases just mentioned, an experience that is “obscured from collective understanding” is assimilated to a different one. To spell this out a little further, consider ‘postpartum depression.’ Suppose, as is plausible, that Sanford was already familiar with some significant normative properties of depression, such as that sufferers ought not be blamed for certain behaviors. In that case, seeing what is happening to her as a kind of depression allows Sanford to recognize those properties of her own experience. That looks like an achievement. Moreover, it's an achievement even if the normative contours of the more familiar experience aren't themselves perfectly understood. This was arguably the case for Sanford: at the time, depression was (even) more socially stigmatized than it is now, and so regarded by many as something to be embarrassed about. Nevertheless, Sanford was still in a better position with the label ‘postpartum depression’ in hand than without it, both in terms of making sense of what was happening to her, and making that communicatively intelligible to (some, though perhaps not all) others.Footnote 12 Similar points can be made about ‘statutory rape’ as well.
This third proposal has several virtues. First, as I've just been arguing, it accommodates the central cases. Recall in this connection the passage from Brownmiller in which an activist (Sauvigne) mentions several options for naming Wood's experience; most of these – ‘sexual intimidation,’ ‘sexual coercion,’ ‘sexual exploitation on the job’ – were quickly rejected. That discussion fits well with the current proposal. For one thing, it's clear that the participants in that discussion were indeed searching for a suitable label. Further, the current proposal can help explain why the alternate labels might have been rejected. ‘Exploitation on the job,’ for example, is not a normatively familiar category, nor does it clearly fit all the cases in question.
Next, unlike the second proposal, the current proposal captures a distinctive injustice, not one that is ubiquitous. Third, and relatedly, it explains how the experience in question can be made communicatively intelligible, and so, how a primary harm of the injustice can be overcome. Fourth, it explains the ‘Eureka’ moments in the first two cases (the Sanford and Wood cases). That is, it explains how acquiring a suitable label can itself be illuminating, and so, how progress can be made in the course of a single conversation.
The third proposal has the following interesting implication: the prospects for overcoming hermeneutical injustices of this sort will be constrained by the normative resources available in the context. As is suggested in the passage from Brownmiller discussed just above, the lawyers and activists thinking about Wood's experience have to choose from labels that already have some currency in their interpretive context. I'll say more about this consequence in the concluding section.
A downside of this proposal is that it doesn't accommodate all the cases of hermeneutical injustice mentioned in the literature. In particular, it doesn't explain why rape myths and domestic violence myths should count as hermeneutical injustices; though I won't argue this in detail here, it also fails to fully accommodate white ignorance as a hermeneutical injustice. However, as suggested in my discussion of the previous proposal, there are important differences between these cases and my central cases. Even if we think that these are all ultimately hermeneutical injustices, it's still worth tracking the differences between them.
Finally, recall Fricker's suggestion that overcoming hermeneutical injustice constitutes not just an achievement, but a cognitive one. The current proposal certainly makes clear that there are achievements in the central cases: making sense of an experience to the extent of rendering it communicatively intelligible to others is definitely an achievement, one with important political consequences. But is this a cognitive achievement? In the remainder of this paper, I'll show why this turns out to be a difficult question.
4. CASE STUDIES
To better understand what is involved in overcoming the hermeneutical injustices in my central cases, it will help to look more closely at how the labels at issue there actually function. In this section, I'll focus on two labels: ‘statutory rape’ and, more briefly, ‘sexual harassment.’ I'll argue that these labels distort our understandings of the underlying experiences. As we'll see, however, there are different kinds of distortion involved in the two cases.
To set up that argument, though, it will help to begin with a brief detour through some recent discussion of a very different label: ‘intellectual property.’ Some have argued that we should reject this label, because it distorts our understanding of the entities to which it applies, and thereby leads us to make bad laws and policies with respect to those entities. I won't take a stand here on whether these conclusions are correct. Rather, I'll use this discussion to illustrate what I mean by ‘distortion’ in this context, and to introduce a distinction between two kinds of distortion. I'll then put that distinction to use in my investigation of the two labels mentioned in the previous paragraph.
The phrase ‘intellectual property’ applies to entities of several types, including (but not limited to) copyrights, patents, and trademarks (WIPO n.d.). Roughly, copyrights cover literary and artistic works, including film, music, and architectural design; patents apply to inventions; while trademarks provide markers to help identify the sources of particular goods and services. Each type of entity is governed by a body of law and policy that, among other things, specifies the rights and responsibilities of owners and relevant others. The label is of relatively recent vintage, and has come to be widely used only in the last three decades.Footnote 13
Criticisms of the label focus on several aspects, of which I'll mention two. First, use of the label suggests a close analogy between the entities mentioned in the previous paragraph, and physical property. But that analogy, it is argued, is flawed (Stallman Reference Stallman2018). Rights to physical property are stable over time, can be bought and sold in a marketplace, can be passed down within the owner's family from generation to generation, and so on. To regard copyrights, patents, and the like as analogous to physical property is to license similar inferences about these entities, specifically, inferences about how rights to the latter should be treated. And while it's easy to see why such inferences may be congenial to the interests of some large corporations like Disney, they are not as clearly in the public interest.
A second line of criticism looks at the differences between the various entities that fall under the label ‘intellectual property,’ and argues that these differences are both normatively significant, and elided by the grouping. For example, a very common argument for respecting intellectual property rights is that doing so encourages innovation (see, e.g., WIPO n.d.: 3–4). But even if that's a good argument for certain types of intellectual property, other types – e.g., trademarks – have very little to do with innovation (Stallman Reference Stallman2018). Trademarks help us identify the producer or seller of a given product; precisely the same stuff, with no innovation to distinguish it, may be sold by other sellers, under distinct trademarks. Failure to recognize this difference between trademarks and, say, patents encourages us to regard all intellectual property as aiming at innovation, and to make laws and policies accordingly.
These critiques suggest that use of the label ‘intellectual property’ distorts our understanding of copyrights, patents, trademarks, etc. It does this in at least two different ways. First, it suggests an analogy, between intellectual property and physical property, that licenses bad inferences about the former. Call this an ‘external distortion’. Second, it groups together entities in a way that elides crucial normative differences between them, or suggests a normative commonality among them that is in fact illusory, or both. Call this an ‘internal distortion’. Both kinds of distortion encourage the creation of laws and policies that fail to reflect the public interest. Or so the argument goes.
As mentioned earlier, I don't plan to adjudicate the correctness of these arguments in this paper. Instead, I will use the framework just sketched to think about labels that are of more central concern to our discussion. I'll turn to that below.
One point of clarification before continuing: in this paper, I am primarily interested in our ordinary understandings of statutory rape and sexual harassment, and our ordinary use of the corresponding labels. Of course, in both cases, the ordinary understanding is informed by legal doctrine (and in the case of postpartum depression, by medical doctrine). And it is a difficult task to separate these aspects of the various labels. Nevertheless, insofar as this separation is possible, my focus is on the ordinary understandings and usage, not the legal technicalities.
4.1 Statutory rape
Let's start with ‘statutory rape.’ Recall that I argued earlier (in Section 2) that the adoption of this label can plausibly be regarded as overcoming a hermeneutical injustice, much like ‘sexual harassment’ and ‘postpartum depression.’ In this section, I'll argue that the phrase also brings with it significant costs: more specifically, in the terminology just introduced, it brings with it both an external distortion, and an internal one.
First, ‘statutory rape’ suggests a close analogy with non-statutory rape. But that analogy is seriously flawed. To see this, we can start with the observation that non-statutory rape is popularly understood as non-consensual sex (though see MacKinnon Reference MacKinnon2016 for trenchant criticisms of that construal). Many instances of statutory rape clearly feature consent, on any reasonable understanding of that notion. Consider, for example, two 15-year-olds engaged in a loving, stable, non-coercive sexual relationship. Their activity will count as statutory rape in many states, particularly those lacking age-span provisions.Footnote 14 Nevertheless, it is bizarre to think of this as rape, understood as non-consensual sex.Footnote 15 It is also bizarre to infer that the 15-year-olds, if convicted, should be treated as rapists, such as by being placed on sex offender registries. If that's right, then the analogy with non-statutory rape creates an external distortion.
Note that the argument above makes no assumption about whether sexual activity between 15-year-olds is – morally, developmentally, or in other ways – appropriate. Regardless of whether one finds that sexual activity troubling (or not), the point above is only that the category ‘rape’ seems ill-fitting in cases like the one described. Further, the argument takes no position on whether the use of sex offender registries, or any other particular treatment of those convicted of rape, is justified. To say that the label ‘statutory rape’ creates an external distortion is, in part, to resist the implication that offenders should be treated in the same way as those convicted of non-statutory rape, whatever that treatment happens to be.
A not unusual response to the argument above emphasizes the distinction between consent in the moral sense and consent in the technical, legal sense. On this line of thinking, sexual intercourse between the 15-year-olds in a loving, stable, non-coercive relationship features consent of the former type, but not of the latter; so it is (only) technically rape. But it is hard to know precisely what this means. Worse, this claim brings with it a different unattractive consequence. To say that statutory rape is only technically rape is to license (further) problematic inferences about its seriousness: specifically, it suggests that instances of statutory rape generally don't constitute serious wrongs, or perhaps wrongs at all. And that encourages us to be cavalier about charges, or convictions, for statutory rape.
Second, ‘statutory rape’ also groups together conduct that is significantly different, normatively speaking. Consider here the difference between the sexual conduct of the two 15-year-olds mentioned above, and a violent sexual assault carried out by a 30-year-old on a 13-year-old. To consider both of these statutory rape is to say that same kind of wrong is done in each case, perhaps with a difference in degree. That again seems like a bad result. Even if there is some wrong done in each case, it's a mistake to think that it's the same kind of wrong, or even similar enough to be treated in the same way. To see this, note that the second case, but not the first, centrally involves a grievous denial of autonomy. Grouping the two cases under ‘statutory rape’ obscures this important normative difference, and encourages us to treat violations and violators in similar ways. If that's right, then this grouping creates an internal distortion, in addition to the external one discussed earlier.
It may be objected at this point that, even if I'm right about the distortions just described, there is nothing special about ‘statutory rape’ here. Rather, pretty much every label groups together instances that differ in some respects, and in so doing, elides some differences between them. To take an example: ‘Sexual assault’ covers a broad range of instances, from touching to intercourse to other kinds of sexual behavior. But use of that phrase doesn't mislead us into reacting similarly to all cases. And the same goes for many other labels in common usage.
There are several points to make in response. First, even if it's true that all groupings elide some differences, not all elisions are equally significant. That's because some differences are more significant than others, and a label that elides more significant differences is less useful than another that elides less significant ones. Second, I take it that an important part of the point of normative language is to help us understand the nature of wrongs, and to direct us towards remedies. Therefore, from the point of view of assessing such language, differences between wrongs (or appropriate remedies for wrongs) are more significant than other differences. A label that collects together, say, premeditated murder and toe-stubbing is unlikely to be a useful piece of normative language.
Of course, the functions of normative language mentioned above can come apart. Some labels collect together behavior that constitutes the same wrong (or similar enough ones), but merits different responses. ‘Sexual assault’ is arguably an example of this type: different types of sexual assault may require different remedies, but it's at least plausible that all types involve the same wrong (to different degrees). Other labels collect together behavior that constitutes different wrongs that nevertheless merit similar treatment, for deterrence or other reasons. A label that collects together the conduct of the perpetrator of some wrong and a (mere) co-conspirator may belong here, insofar as we want to punish both equally for deterrence purposes though we recognize that they've committed different wrongs.Footnote 16 But the mere fact that a label collects together instances with differences among them isn't enough to show that it is distorting in the sense relevant to this paper.
Finally, and most importantly for my purposes, what I've been arguing in this section is that ‘statutory rape’ fails to fulfill either of the functions of normative language just described: it neither collects together behavior that constitutes similar enough wrongs, nor behavior that merits similar treatment. It thus clearly fails to achieve the very point of normative language, unlike the labels discussed in the previous paragraph.
4.2 Sexual harassment
I'll turn next to consider one final example, namely, ‘sexual harassment.’ In what follows, I'll sketch an argument to the effect that this label creates an internal distortion, by suggesting a normative commonality that is in fact illusory. My conclusions in this section will be somewhat tentative; I hope to return to this topic in future work.
As we've already seen (Section 2), our current understanding of sexual harassment owes much to the work of feminist lawyers and activists in the 1970s. Catharine MacKinnon's work on this topic was particularly influential. MacKinnon argued that sexual harassment in the workplace constitutes sex discrimination, and as such, should be prohibited under the employment discrimination provisions of Title VII of the 1964 Civil Rights Act (MacKinnon Reference MacKinnon1979). She further distinguished two types of behavior that should count as sexual harassment, namely, quid pro quo sexual harassment and hostile environment sexual harassment. Both types were subsequently recognized under the guidelines issued by the Equal Employment Opportunity Commission (EEOC).
The current EEOC guidelines characterize sexual harassment as “[u]nwelcome sexual advances, requests for sexual favors, and other verbal or physical conduct of a sexual nature” that is made “either explicitly or implicitly a term or condition of an individual's employment,” and that falls into one of the categories below:
Quid pro quo: Submission to the sexual conduct is “used as the basis for employment decisions affecting such individual.”
Hostile environment: The sexual conduct has “the purpose or effect of unreasonably interfering with an individual's work performance or creating an intimidating, hostile, or offensive working environment” (EEOC 1999).
Of these, hostile environment sexual harassment was by far the more controversial category. The central issue in the first Supreme Court case on sexual harassment (Meritor v. Vinson) was whether harassment of this sort was genuinely prohibited by Title VII; overturning lower court rulings, the Court held that it was.
There are several important differences between quid pro quo and hostile environment sexual harassment. For one thing, as the label suggests, quid pro quo sexual harassment requires that some exchange or trade be demanded, while hostile environment sexual harassment does not. Further, quid pro quo sexual harassment also requires that some “tangible” economic benefit – such as job retention, or salary, or promotion – be affected, while hostile environment sexual harassment does not. Quid pro quo sexual harassment typically involves an agent (or a few agents) unjustly using their social power to harm another agent (or agents), and as such, can be understood as agent oppression (see Haslanger Reference Haslanger2012a for the distinction between agent oppression and structural oppression). Hostile environment sexual harassment, by contrast, may involve a pattern of behavior, no single instance of which is particularly problematic, but which together ends up “unreasonably interfering with an individual's work performance”; as such, it may be better understood as structural oppression. And there are further differences as well.
These differences already suggest that ‘sexual harassment’ collects together quite a range of behavior. But as noted above (Section 4.1), that by itself isn't enough to sustain a charge of distortion. However, the differences mentioned in the previous paragraph do at least raise the following question: in light of these differences, is there good reason to group together the types of behavior that ‘sexual harassment’ collects, while excluding other types, including other types of harassing behavior? Let's call this ‘the commonality question’.
To see why this might be a pressing question, note first that ‘sexual harassment’ seems to exclude several further types of harassing behavior that are not (or not always) sexual, but that regularly accompany paradigmatic instances of sexual harassment. Elizabeth Anderson offers several examples of such behavior, including all of the following:
Gender harassment, which includes co-workers expressing hostility towards women doing ‘men's work’, or acting in ways that make it impossible for women to function in male-dominated workplaces.
Gender policing, which involves harassment of gender-nonconforming individuals.
Sexual orientation/identity harassment, which includes taunting, bullying, and assault of LGBTQ individuals.
“Horseplay,” where male harassers engage in bullying and hazing of male coworkers (Anderson Reference Anderson2006: 287–8).
The types of behavior on this list relate to gender, gender identity, or sexual orientation broadly speaking, but needn't take a sexual form. For example, co-workers telling a woman firefighter that women can't do the job may qualify as gender harassment, but isn't particularly sexual at all. Nevertheless, these types of behavior are important to include in any discussion of sexual harassment because they often occur together. In fact, it may be difficult to understand the full scope of the problem in a particular workplace without paying attention to all of these types of behavior as well.
I've argued thus far that ‘sexual harassment’ collects together quite different types of behavior, but excludes other types that are very closely related. Nevertheless, there might still be a satisfactory answer to be given to the commonality question if, for example, it could be shown that there is some distinctive wrong of sexual harassment. This is a difficult question that cannot be fully settled here; but I'll end this section by suggesting some reasons to be pessimistic about the prospects of a satisfactory answer.
As we saw above, MacKinnon thinks that the wrong of sexual harassment is sex discrimination. She further offers two approaches to understanding sex discrimination, and therefore, two ways of unpacking the wrong of sexual harassment (MacKinnon Reference MacKinnon1979, especially Chs. 5–6). Nevertheless, the view of sexual harassment as sex discrimination faces a serious challenge from cases of same-sex sexual harassment. MacKinnon anticipates this problem, and offers some resources for addressing it. For example, on one of the approaches she describes (the “inequality approach”), sexual harassment consists in the harasser, playing a masculinized role, rendering the harassee as sexually subordinate, by putting the latter in a feminized role; this allows the harasser and harassee to be of the same sex. Still, as Anderson argues, even if this move can accommodate some cases of same-sex harassment, it's doubtful that it can accommodate all cases (Anderson Reference Anderson2006: 309).
Other theorists have offered alternate accounts of the wrong of sexual harassment. Arguably, these do even worse in characterizing this wrong in a way that gives a satisfactory answer to the commonality question. Consider the view of sexual harassment as a dignitary affront (Brooks Reference Brooks1999, among others). While this may well capture all cases of sexual harassment, it seems to also capture several of the types of behavior on Anderson's list. As such, it also doesn't offer a satisfactory answer to the commonality question.
There is, of course, more to say here. In particular, there are further accounts of sexual harassment that I haven't discussed in this section. Nevertheless, what I have said is enough, I think, to suggest that an adequate answer to the commonality question will not be forthcoming. If that's right, then ‘sexual harassment’ creates an internal distortion, by suggesting a normative commonality that is in fact illusory.
5. TAKING STOCK
In this paper, I've argued for two main conclusions. The first of these is a characterization of the gaps in some central cases of hermeneutical injustice. On my view, what's missing in each of these cases is a label that assimilates the experience in question to others with which we are already normatively familiar. The second main conclusion is that overcoming hermeneutical injustice via the adoption of such labels will regularly involve introducing distortions into our understanding of the experiences. I've illustrated this via discussion of two examples.
Two caveats are in order here. First, to say that overcoming hermeneutical injustice regularly requires introducing distortions is not to say that anyone consciously chooses to distort their understanding of the experience at issue. Rather, as I hope is illustrated by the examples in the previous section, the distortion is a result of the constraints imposed by the already available linguistic/conceptual resources. Second, my argument isn't intended to show that overcoming hermeneutical injustice – even in cases like my central ones – always involves introducing such distortions. For all I've said, perhaps the gaps I'm concerned with are sometimes filled without there being any distortions at all. Consideration of further examples is needed to settle this issue.
The argument in this paper does put us in a position to draw some lessons. First, we can now say more about the harms of hermeneutical injustice. As we saw in Section 2, Fricker argues that a primary harm of this injustice is that it keeps agents from making communicatively intelligible experiences that they have a strong interest in so rendering. Our discussion here suggests a further harm: in many cases of hermeneutical injustice, agents suffering the injustice will find themselves in the position of having to choose between rendering their experiences communicatively intelligible, and genuinely understanding those experiences. This looks like a double bind in the sense described by Marilyn Frye: either option brings with it significant costs (Frye Reference Frye1983: 2).
A second lesson is more tentative. We saw at the end of Section 3 that overcoming the injustice in my central cases, by making a previously unintelligible experience communicatively intelligible, constitutes an achievement. Whether it is also a cognitive achievement is, at best, unclear. In cases where the distortions introduced are especially significant – as is the case for ‘statutory rape,’ at least – there may be good reason to say that the achievement is not cognitive, but of a more political or practical nature. Which distortions qualify as significant enough in this sense is also a question that deserves further exploration.Footnote 17
