Over-intelligibility

I. CASE STUDIES

This section presents three cases of over-intelligibility. The first concerns concepts of sexual identity, the second concepts of sexual violence, and the third concepts of race. While these cases can be read in multiple ways, I will argue that one plausible reading is that they show that intelligibility can have epistemic downsides. The rest of this article will then explain how to develop such a diagnosis.

II. ANTI-INTELLIGIBILITY

Conceptualization can backfire. A version of this position, closely associated with the work of Judith Butler and Michel Foucault, was prominent in continental philosophy and critical theory circles from the 1970s through the 1990s. Butler and Foucault are often associated with opposition to treating certain kinds of categories (e.g. gender categories or mental illness categories) as natural, ahistorical, or having an essential nature. However, their respective oeuvres also defend the position that conceptualization or intelligibility can be inherently subordinating or dominating. This section examines those positions. It argues that the most plausible ways of spelling out such positions risk collapsing into the familiar point that the truth can be practically disadvantageous.

In The Psychic Life of Power, Butler speaks of the “self-colonizing trajectory of certain forms of identity politics.”³⁹ For Butler, this turns on a kind of problem that accompanies the subject’s categorization: “Social categories signify subordination and existence at once.”⁴⁰ Butler justifies the claim of “subordination” by claiming that the subject is “[b]ound to seek recognition of its own existence in categories, terms, and names that are not of its own making.”⁴¹ On Butler’s view, recognition is required in order to be a subject, and recognition occurs in accordance with norms (e.g. gender norms), which in turn invoke categories (e.g. man/woman). So the overall thought is that, in being recognized, someone is rendered intelligible in a way that dominates them, insofar as the categories that figure in that act of recognition have their origin external to the subject. But is it really a problem if such categories are not of my “own making”? This seems to presuppose too strong a standard of what it takes to be non-subordinated.

A different way of developing Butler’s argument turns on the point that subjects seek recognition through discourses that are “dominant and indifferent.”⁴² That is, the categories through which we recognize one another in everyday contexts, or even putatively liberatory contexts, have histories of domination: the sexological categories of 19th-century medical discourse, the gender categories of heteropatriarchy, the racial categories of 18th-century racial science. Perhaps, in being thus categorized, we are playing into the hands of these oppressive discourses. This is one way to understand Butler’s account, in Gender Trouble, of “the category of women” as it functions within feminism to define a group of “political subjects,” which on Butler’s view repeats the patriarchal pretense that sex is the stable underpinning to gender.⁴³

But this falls short of telling us about drawbacks inherent to intelligibility. Recognizing someone as a woman and thereby as a “subject” of feminism may be an intelligibility-conferring practice, but the problems Butler identifies with that practice concern the way woman functions within it, not intelligibility itself.

Foucault’s work provides another influential reference point for critiques of intelligibility. This is most directly suggested by Foucault’s account of the interrelation of power and knowledge; for instance, his claim, in Discipline and Punish, that “power and knowledge relations… invest human bodies and subjugate them by turning them into objects of knowledge.”⁴⁴ Foucault thereby suggests that rendering-intelligible is itself a way of subjugating. Likewise, in The History of Sexuality: Volume I, Foucault speculates that a proliferation of different sexualities is not a rebellion against power. Rather, the different “species” of sexuality are the effect of a “machinery of power” that uses a “principle of classification and intelligibility” in order to “strew reality” with such “species” and “incorporate them into the individual.”⁴⁵

Here intelligibility itself is understood as a dominating force. But it is hard to see how to interpret this epistemic act or relation as one of subjugation or domination. Instead, Foucault has commonly been understood as pointing out how intelligibility makes possible subsequent acts of subjugation. For instance, Linda Martín Alcoff ascribes to Foucault the view that “bringing things into the realm of discourse can create new opportunities for discipline and normalization.”⁴⁶ One example would be categories of sexual acts, such as sodomy, whose articulation as a distinct category renders criminalization possible. But in this case, it seems that the relevant problem is not intelligibility itself, but rather the unjust legislation that intelligibility makes possible. Moreover, it seems that this strategy ultimately identifies downstream practical consequences of intelligibility, leaving us with a version of the familiar point that knowledge can be disadvantageous.⁴⁷

Finally, Foucault’s account of the “power of normalization” suggests the possibility of drawing a link between intelligibility and normalization.⁴⁸ For instance, the examination, which “makes it possible to qualify, to classify and to punish” enables the school to assess its pupils in order to determine which students diverge from expected standards, which then enables penalties to be assigned so as to motivate the students to become more normal or similar to one another.⁴⁹ As noted previously, it does not seem that the problem here is with intelligibility, but rather with a social process that intelligibility facilitates.

Some Foucaultian thinkers have made the stronger claim that intelligibility itself normalizes. For instance, Wendy Brown argues in States of Injury that the “discourse of politicized identity” reiterates the “normalizing principle of disciplinary regimes.”⁵⁰ How might this work? Elsewhere, by way of example, Brown claims that the effort to articulate “the feminist truth about women” creates a “unified discourse” in which “the porn star who feels miserably exploited, violated, and humiliated in her work invariably monopolizes the feminist truth about sex work.”⁵¹ Brown also notes that the “putting into public discourse of heretofore hidden or private experiences” can bring about “homogenization.”⁵² These suggestions offer a promising avenue for making sense of the previous section’s cases. For one can understand Sontag as resisting the “homogenization” that a vocabulary of sexual identities would impose on her erotic life, or Maria as recalcitrant to how an activist discourse might “monopolize” the truth of bad sexual encounters. But we need a more precise account of how, exactly, such homogenization or monopolization takes place. Why should such dire consequences be associated with the mere application of a concept such as bisexual or rape? Brown claims that “speech, because it is always particular speech, vanquishes other possible speech.”⁵³ But this explanation risks looking like hyperbole. Although these concepts do not capture the full complexity of what they describe, as noted previously, we do not in general resist using words (such as “love”) that are similarly incomplete when used of their objects.

This article agrees with Butler and Foucault that intelligibility has downsides, and moreover, downsides grounded in its nature as an epistemic good. Moreover, it agrees with Foucault that “normalization” is a crucial concept for understanding intelligibility’s downsides, and with Brown that these downsides are closely linked to dynamics of monopolization and homogenization. But this article also seeks a more rigorous explanation for these points. Butler, Foucault, and Brown are right to be suspicious of intelligibility, but we need a better account of why they are right.

III. NORMALIZING SALIENCE

This section explains one way intelligibility has epistemic downsides: by standardizing knowers’ lives in a way that is inapt, given their epistemic ends. First, I introduce the scaffolding for this diagnosis. Concepts modulate salience, or affect which properties are particularly prominent, and therefore available to figure in inquiry and interpretation. Assignments of salience are apt if they facilitate a particular knower’s epistemic ends. Second, I argue that concepts can bring about normalization, or epistemically bad standardization, in the following way. How concepts modulate salience reflects collectively shared ends. This means that by deploying the same concepts, we become more similar with respect to what we treat as salient. This standardization of salience is epistemically bad for a knower just in case a concept’s way of modulating salience is not apt, given that knower’s epistemic ends.⁵⁴ In such cases, there is normalization. Third, I argue that this normalization, which arises from a mismatch between a knower’s ends and a concept’s way of structuring salience, should not be analyzed as a case of incomplete or imperfect intelligibility.

Concepts not only set forth necessary conditions for their application, but also affect which features are rendered salient, or particularly available to feature in “cognitive processes like memory, categorization, recognition, and even decision-making.”⁵⁵ What is salience? This is a difficult notion to define;⁵⁶ I will spell out what I mean by drawing on Elisabeth Camp’s work on characterizations.⁵⁷ Characterizations structure our thinking by rendering some features of an object more prominent than others.⁵⁸ Having a characterization is not a matter of endorsing certain “thoughts about the prominence, centrality, or fittingness of the characterization’s constituent features.”⁵⁹ Rather, a characterization “actually structure[s] one’s thoughts” in a way that makes certain properties readily cognitively available to figure in further inferences or interpretation.⁶⁰ For instance, my characterization of a restaurant renders serving food prominent, or particularly likely to jump to mind when I think of restaurants. This explains why, when I walk past a building on Main Street that I recognize as a restaurant, I am more likely to attend to a menu posted on the window, or wonder if I might apply for a job there, than I would be if I simply characterized it as a building. Concepts not only express necessary conditions but also “characterize” or modulate salience (i.e. prominence), which affects the availability of an object’s properties for inference and inquiry.⁶¹

Characterizations form parts of perspectives. A perspective, as Camp defines it, is a disposition to characterize.⁶² Camp’s examples include political liberalism, political conservatism, and evangelical Protestantism.⁶³ Rachel Fraser, building on Camp’s work, glosses a perspective as a “way of looking at the world” constituted by a “suite of interlocking dispositions.”⁶⁴ Such dispositions are “attentional” (that is, dispositions to direct one’s attention in certain ways), “inquisitive” (suggesting which modes of inquiry are worthwhile), and “interpretive” (disposing one to make certain inferences and evaluations) (Fraser p. 4028). In the remainder of this article, I will use Fraser’s account of perspectives. If I have a feminist perspective, then in observing the dynamics of my classroom, I am disposed to attend to whether the male students are paying equal attention to female students’ points. Upon reading in the newspaper that the government has passed a new policy to expand healthcare provisions, I am disposed to ask what kinds of reproductive care are included. The way my concepts characterize subject-matters form parts of dispositions to attend, inquire, and interpret that make up my perspective(s).

Both characterizations (i.e. ways of structuring what is prominent or salient), and the related dispositions to attend, inquire, and interpret, can be evaluated as epistemically good or bad, depending on whether or not they are apt.⁶⁵ Camp holds that our characterizations are better or worse depending, in part, on whether they are “conducive to achieving one’s current cognitive goals.”⁶⁶ Conduciveness, or aptness, can be understood as consisting in how well or poorly a characterization facilitates an epistemic end. This introduction of aptness, given our ends, implicitly appeals to our finitude, or situation as non-ideal epistemic agents.⁶⁷ There are many more facts to be known than any human knower can ever know; more interpretations to be made than any knower can ever make; more understandings to be achieved than any knower can accomplish. We have finite resources and therefore cannot attend to every property, ask every question, or achieve every interpretation. This is why Friedman argues that our interests play a role in determining which justified beliefs we ought to hold.⁶⁸ A characterization, salience-affordance or disposition to inquire, interpret, and attend is apt if, for a given agent with a set of epistemic ends, that characterization, affordance, or disposition is conducive toward achieving that agent’s epistemic ends. It is inapt if it is not conducive toward achieving that agent’s epistemic ends.

This can help us make sense of Sontag’s resistance to being called bisexual or a lesbian. Consider the following example. An entry from the 16-year-old Sontag’s diary recounts a period she spent living in the Berkeley student dorms. This entry moves though a series of topics: the “dichotomy between the body and the mind” that Sontag describes as having been her “greatest unhappiness,” a lecture on aesthetics, her relationship with a man (Allan), a subsequent relationship with a woman (Harriet), a proclamation that she “want[s] to sleep with many people—I want to live and hate and die,” which she connects with her desire not to become an academic (“I don’t give a damn for anyone’s aggregation of facts”), scenes from gay nightlife in San Francisco with Harriet, a list of classical music titles followed by “Sex with music! So intellectual!!”, notes on Goethe and Myshkin, and a list of “gay slang.” In this entry, Sontag refers to her “lesbianism” and “my homosexuality.”⁶⁹ What might the stakes have been, for the later Sontag, of declining to use these labels?

One point is that, in thinking of herself as a lesbian, when Sontag reflects on her own erotic history, the details relating to gender and same-sex life in San Francisco in 1949 may jump to mind more readily than details about the relationship between mind and body. In thinking about her relationship with Harriet, the role that Harriet played in introducing Sontag to San Francisco gay life might be more prominent in Sontag’s thinking than their shared intellectual life (“I read the first pages of her copy of Steppenwolf…”).⁷⁰ Properties that are relevant to interpreting Sontag’s time in Berkeley as a narrative of gay awakening become salient or prominent; properties that are relevant to interpreting it as part of an autobiography of artistic awakening become less prominent. This point can explain why Sontag would have an epistemic reason to resist thinking of herself as a lesbian or bisexual: using those concepts structures what is salient in Sontag’s mental life in a way that can be inapt given her epistemic ends. In particular, focusing on gendered relationship dynamics can be inapt for a ruminative aesthete who links her erotic life with its “spasms of uncontrollable desire” to the “contemplation of infinity” (and who writes after this entry, “How dreary and monotonous [these notebooks] are!”).⁷¹ While the belief “I am bisexual” is not a junk belief in Friedman’s sense, it has what we might call a junky epistemic profile for Sontag, insofar as it renders gender-and-sexuality properties salient at the expense of others.

A similar point can be made about Maria’s first sexual encounter. The characterization of rape plausibly involves treating some of an event’s features as prominent, such as those involving failure of consent (e.g. “[s]he said, maybe, maybe and he said, yes, yes”), and others as less prominent, such as those inviting comparison to more prosaic events (“[i]t was gross”). Treating the failure of consent as prominent is not a matter of forming a belief about how much the consent-failure matters, but rather consists in the feature’s being particularly prominent, and hence highly available to serve in inquiry and interpretation.⁷² This structuring of salience is apt if Maria is writing an essay on statutory rape. It is inapt if Maria is reflecting on this event as potential material for, say, a dark comedy with body-horror elements.⁷³ In this latter case, a more apt characterization of the event would render salient the “gross” elements of that encounter—smells and textures, facial expressions, and noises. This does not mean that Maria would form false beliefs, or worse interpretations, if she were to use the concept rape. Rather, she would have been more likely to form different beliefs, and different interpretations.

If Sontag or Maria were to use the concepts offered to them, what would result is normalization, or epistemically bad standardization. To see the link to standardization, we should first note that the salience-profiles of concepts reflect the epistemic ends of broader groups or collectives. To label an event as genocide or sexual harassment or rape, or to label a person as lesbian or counter-cultural or mentally ill, is to render salient certain features of that event or person which are apt relative to the aims of different groups, for instance: victims, policymakers, feminists, doctors. The ends adopted by members of groups are facilitated by putting forward conceptual resources that promote shared attention to certain properties at the expense of other properties. Such conceptual resources thereby promote what Fraser calls “perspectival co-ordination,” through which knowers converge on how they attend, inquire, and interpret.⁷⁴

The concept rape is standardizing, for Maria, insofar as it makes her mental life more similar to those of the “women in her self-defense class,” and make her more likely to coordinate her perspective (dispositions to attend, inquire, interpret) with theirs. If she were to repeatedly do this, in the long run their beliefs or opinions would be more likely to converge. Fraser notes that “perspectival co-ordination” can generate what she calls “opinional convergence,” insofar as “agents who structure information in the same ways are far more likely to remember the same things, and to draw the same inferences, than those who structure the same information differently.”⁷⁵ Likewise, the concepts bisexual and lesbian are standardizing, for Sontag, insofar as they structure her way of attending to her erotic history in a way that is coordinated with either activist understandings or socially standard scripts for thinking about non-heterosexual lives. Since concepts not only articulate necessary conditions, but also structure salience, they promote collective convergence around salience, or make us more similar with respect to what we treat as salient.

The concepts lesbian and rape standardize. But why think that this is epistemically bad standardization, hence normalization? Not all standardization is epistemically bad. When the newspaper reports that there will be a thunderstorm today, its readers converge on believing that there will be rain, and hence become standardized, or more similar, with respect to their rain-beliefs. Standardization can be epistemically bad when a knower’s mental life is standardized in a way that is inapt given their epistemic ends. Focusing on consent-failure could be inapt for the writer of darkly comedic body horror; focusing on gendered relationship dynamics could be inapt for the ruminative aesthete. Intelligibility, through concepts that would enable Maria and Sontag to believe truths and gain understanding, is simultaneously a vector of normalization. This explains how intelligibility can have epistemic downsides.

Now, the overintelligibility-skeptic might present the following objection to the line of thought developed thus far. If there is an epistemic problem with bisexual or lesbian or rape of the general sort just described, this might seem to be a matter of holding false beliefs, or faulty understanding: for Sontag, inaccurately interpreting the time in Berkeley; for Maria, inaccurately interpreting her first sexual encounter. In that case, it would seem that the problem really is a failure of intelligibility.

This counter-argument goes wrong in supposing that, when two interpretations differ, one must be superior with respect to intelligibility. But for finite knowers such as ourselves, we can face situations in which distinct interpretations are on par with respect to intelligibility, and so our reasons to prefer one are not explicable in terms of gains of intelligibility. Consider Sontag’s time in Berkeley. Rendering gender-and-sexuality-related properties of that experience salient may be conducive to the epistemic ends of, say, a historian writing about non-heterosexual celebrities of the 20th century. For such a historian, foregrounding those properties of Sontag’s own experience would be apt. But for an aesthete such as Sontag, it is not. Different ways of structuring salience are more or less suited to different projects. An account of Sontag’s time at Berkeley that shows how Sontag partook in a mid-century gay experience, or an account of that same time that focuses on what Sontag was reading as part of her aesthetic self-development, can be different but equally intelligibility-conducive interpretations, suited to different projects. For finite knowers, multiple ways of structuring salience can lead to equivalent intelligibility, and we should prefer some over others because they are more apt given our projects or ends.

Over-intelligibility obtains when a concept structures salience in a way that generates intelligibility, but is inapt given a particular knower’s ends. It might seem, however, that this diagnosis falls short of explaining Sontag’s horrified reaction to her interviewer. Why could Sontag not have conceded that some people might choose to use the label “bisexual” or “lesbian” about her, but then simply note that she herself does not find those labels helpful? Sontag could have accepted the concepts’ application to her while declining to structure her own inquiry and interpretation using those concepts.⁷⁶

This strategy can mitigate some of the effects of over-intelligibility in some circumstances. However, as a blanket proposal, it underestimates the social and collective nature of inquiry in general, and end-setting in particular. Having told other people that they might label me as a lesbian, I am more likely to face queries from them about this aspect of my erotic life: “Are you out to your colleagues?” “Is that why you and Harriet were always hanging out together?” I could try to establish a separation in my own mind between others’ interpretive projects, and my own. But the more frequently other people use a certain concept when interacting with me, the more I must use it myself. This naturally leads to the shaping of my attentional, inquisitive, and interpretive dispositions.

Moreover, if a concept and associated set of inquisitive and interpretive ends is current in my social environment, then if I choose to prescind from the associated epistemic resources, I risk estranging myself from others. This estrangement is not merely an emotional experience. When I share my epistemic ends with other people, their inquiries are helpful to me, and vice versa. When my epistemic ends are relatively idiosyncratic, I am less likely to be able to use others’ investigations to facilitate my own. Popular uptake of particular concepts can be associated with a large amount of resources for inquiry, interpretation, and attention being collectively expended on the ends relative to which those concepts are apt. This, in turn, leaves fewer of those resources available for alternative projects.

The availability of concepts such as bisexual and lesbian contributed to an increasing number of concept-users adopting epistemic ends that are facilitated by those concepts’ salience-profiles. This, in turn, created an interpretive culture in which Sontag faced continual scrutiny over the genders of her sexual partners. This corresponds to an allocation of relatively more epistemic resources to sexual-identity-focused projects (such as biographies and articles about Sontag that examine her denial of her putative sexual identity), and relatively fewer resources to other projects that Sontag thought were more worth pursuing. Sontag’s resistance was partly a matter of struggling with others in her social environment over the very question of what questions are worth asking.

The diagnosis of over-intelligibility suggests a shortcoming in projects of hermeneutical justice. When articulating new concepts, we not only make it possible to apprehend new truths, but also promote collective convergence around salience. That collective convergence can be helpful for political projects. But it can also disadvantage knowers trying to engage in idiosyncratic inquiries into that same domain, not only because this leads to pressure to inaptly standardize, but also because there is a reduction in the collective interpretive and inquisitive resources available for idiosyncratic projects. This is one reason intelligibility can be bad: while intelligibility consists in truth and understanding, it can also detrimentally standardize inquiry, attention, and interpretation. Articulating new concepts thus leads not only to intelligibility, but also to over-intelligibility, via an epistemically bad standardization of cognitive life, or normalization. This section has developed the diagnosis of normalization with reference to a mismatch between particular knowers’ epistemic ends and the salience-profile of a concept. The next section develops a second kind of diagnosis that investigates how normalization affects the ends that we set.

IV. NORMALIZING ENDS

The previous section’s diagnosis of over-intelligibility treated knowers’ epistemic ends as fixed. Normalization results from standardizing salience in ways that are bad, given a particular knower’s ends. Structuring salience can also influence what ends knowers set. This section argues that a further, distinct kind of normalization arises when knowers’ ends are detrimentally standardized. This can happen in two ways. First, a knower can have a bad lack of diversity in their ends, which leads to their being stuck in repetitive inquiry. Second, a knower can adopt ends because they fail to recognize that they possess the latitude to direct their inquiry, interpretation, and attention in a different manner.

In order to develop this section’s diagnosis, we need to introduce new kinds of epistemic badness. Let us call a condition normalized just in case it is a condition in which there is an epistemically bad lack of diversity in epistemic ends. Intuitively, the thought here is that too much standardization leads to homogenization. Maria’s use of racial concepts can be understood as a case of normalization in this sense. For Maria, concepts such as biracial and new person form part of a perspective on the world that disposes her to interpret and attend to others in light of their engagement with racial roles; for instance, noticing how her sister-in-law plays up an appearance of non-whiteness in order to appeal to white men. None of what Maria notices is false or unimportant. And yet, racial concepts come to dominate Maria’s trajectory as a knower. She notices racialized manipulation in part because she is primed to look for such behavior, which yields beliefs that further justify her commitments to her epistemic ends, in virtue of which she finds more racialized manipulation. There is a self-reinforcing loop between epistemic ends and the dispositions to attend, inquire, and interpret that subserve those ends. Maria’s attention is directed in this self-reinforcing way toward a narrow set of ends. There is a lack of diversity in her ends, and so she is in a normalized condition.

Not all narrowness is bad. The dedicated activist and the monomaniacal botanist alike might be epistemic successes in virtue of fidelity to a limited conceptual repertoire. But Maria’s alienation—her sense that she or those around her are “wooden”—suggests that she is stuck. Her inquiries and interpretations do not seem to develop or deepen over time (in the manner of a successful botanist), nor do they lead to new proposals for collective action (in the manner of a successful activist). Rather, Maria continually notices versions of the same pattern in different contexts, thereby building up an ever-larger web of true beliefs that is focused on normatively and politically important manners, yet nonetheless always leads her to more of the same. In good cases, lack of variety leads to deepening and complicating inquiry into a single area; in bad cases, it leads to repetition. Maria is in a normalized condition because there is a bad lack of diversity in her epistemic ends. Likewise, if Sontag were to shift her epistemic ends in response to pressure from people like Acocella and Moser, a normalized condition could come about, with a bad lack of diversity within her own epistemic life and epistemic life collectively.

I will now introduce a second way to diagnose normalization at the level of ends. This concerns a bad process through which knowers give up, set, or maintain epistemic ends. I will call this end-enforcement. End-enforcement obtains when we adopt ends because latitude on our epistemic lives has been obscured from us. By “latitude,” I mean our aforementioned capacity to justifiably attend, interpret, and inquire differently from one another. End-enforcement is standardizing insofar as knowers who do not recognize that they have latitude are less likely to use it, and hence more likely to become similar to other inquirers and to adopt group and collective norms. End-enforcement is epistemically bad insofar as knowers give up, adopt, or maintain their epistemic ends in part because their latitude in end-setting has been obscured from their view. End-enforcement is therefore an instance of epistemically bad standardization, or normalization.

Many conflicts that seem like conflicts over what the facts are, are actually conflicts about which facts matter, and by extension, conflicts about what ends we ought to have. Sontag’s critics do not (merely) suppose her to be deceived about her sexual orientation; they also criticize her for not having cared more about it.⁷⁷ Had Maria told the women in her self-defense class that she took the fact that “[i]t was gross” to be more salient to her first sexual encounter than the fact that “[s]he said, maybe, maybe,” it is easy to imagine them telling Maria that she’s wrong about what matters. To say to someone, “You’re a lesbian,” or “You’ve been raped,” can be a matter not only of stating the truth, but also suggesting its heightened salience, and thereby also recommending a prioritization of certain ends (e.g. solidarity among sexual minorities, or victims of sexual violence). Exhorting is not inherently impermissible, although it can go wrong if I thereby meddle in your private business. Being voluble about someone else’s sexual identity or sexual experiences can easily go wrong by overstepping boundaries. But there is a more insidious, yet also ubiquitous, way for exhorting to go wrong. This is when exhortation masks its character as exhortation.⁷⁸

Such a mode of proceeding obscures a crucial dimension of choice. To frame a question about, say, sexual identity in such a manner as to suggest that the truth of the concept’s application exhausts the reasons to form or omit a belief, is also to obscure latitude to attend differently to those facts depending on one’s ends. We thereby become more standardized or similar to one another through a deceptive process. This explains what would be bad about Sontag’s giving up her epistemic projects in response to a claim like Moser’s about what is true. For as the previous section argued, Sontag would not thereby achieve greater knowledge or understanding by adopting a perspective more similar to that of Moser, but rather different knowledge or understanding. I now point out that she might acquire that different knowledge under the guise that she is gaining greater knowledge, insofar as she is “coming to see the truth about herself.” There are possible cases in which Sontag would choose to become more similar to a group through believing that there is no choice to be made. This is epistemically bad standardization, or normalization.

It might seem manipulative to obfuscate one’s exhortations and thereby obscure others’ latitude to accept or reject such exhortation. On this basis, we might think that we should separate figuring out what is true, from figuring out what to do. But this is not possible all or even most of the time. We are continually conceptualizing, and hence continually modulating salience, in ways that are more or less apt given our ends. In conceptualizing, I am already taking a stand.

Can we at least be honest with one another, and with ourselves, about that? For instance, Sontag’s critics might reframe their complaint in terms of Sontag’s resistance to becoming the kind of person who would make identity categories central to her thought. But as a general strategy, this is implausible. As Camp notes, the functionality of characterizations and perspectives depends on their ability to guide us in a rapid and semi-automatic manner. This functionality is at odds with knowers continually making the role of their own ends explicit to themselves or to one another. When we use characterizations and perspectives, we engage in a practice that exhibits self-effacing functionality. Matthieu Queloz introduces this terminology to describe practices for which it is a requirement “on the practice’s functionality that participants not be primarily motivated by awareness of that functionality.”⁷⁹ Self-effacing functionality describes the prioritizations and dispositions bound up with characterizations and perspectives. Both deploying and promoting my perspectives and ends works best when nobody realizes that’s what I’m doing. To make explicit that one is prioritizing the fact that a sexual encounter was nonconsensual is to draw attention to how one might attend differently, and hence to call into question one’s own prioritization. This is not to say that it is impossible for people to be more honest or explicit. But end-enforcement is part of the proper functioning of characterizing and deploying perspectives, or part of how structuring salience works.

End-enforcement can also occur purely cognitively, outside verbal exchange. Maria’s use of racial concepts can be understood as this kind of case. No matter the situation she is in, Maria’s racial concepts help her notice something important about the racial dynamics at play. These racialized negotiations really are happening, and they really do matter. But it comes to seem to Maria that her concepts’ aptness is solely a matter of how they yield knowledge and understanding, and not a matter of her freely set epistemic ends. This is because characterizations exhibit self-effacing functionality in thought as well as in speech. In order for characterizing to structure salience in a way that is cognitively efficient, the role of ends in rendering characterizations salient cannot be continually made conscious. Maria does not consciously recognize that she herself has set the ends in relation to which her racial concepts structure salience in an apt manner. Maria is subject to normalization not only because her epistemic life lacks diversity, but also because that lack of diversity comes about through an obscuring of latitude. She is normalized twice over.

Finally, when characterizations or perspectives are based in moral ends, there is a particular kind of inconsistency involved in acknowledging latitude. The women in Maria’s self-defense class would be unlikely to say, “Sexual violence must be a site of collective resistance—depending on your overall profile of interests and ends.” The anti-AIDS activist slogan “Silence=Death” makes precisely the claim that here one does not have the latitude to attend, inquire, and interpret differently.

Moral ends often do not afford us latitude. To take one prominent example, Kant holds that moral ends are unconditional, by which he means independent of our interests or inclinations.⁸⁰ It would be odd to treat someone as having epistemic latitude to attend (etc.) differently, grounded in their choice of ends, when I take them to lack the moral latitude to omit or deprioritize that end. It may be manipulative to obscure how the aptness of attending (etc.) in certain ways is grounded in our ends or interests—that is, in our choices. And yet, on many views it is precisely the nature of the claim that the good makes on us, that we do not have a choice.

In general, obscuring latitude is not a matter of conceptualization failing. Rather, it is part of the proper functioning of our conceptual practices. Structuring salience works better when we are not conscious that salience is being structured. This means that we are the continual subjects of end-enforcement in both verbal exchange and in our mental lives. That is, we continually make claims on one another, under the pretense of making claims about how things are. In that sense, we are continually manipulating one another: treating beliefs like that was rape or this is sin or you are a lesbian as if the reasons to form them are independent of our ends. We proceed as if in doing so, we are simply stating how things are, when in fact we are recommending to one another that we act; that we bring about; that we become.

This section has explained two ways that intelligibility-conducive conceptualization can normalize our ends. On the first proposal, normalization obtains when there is a lack of diversity in epistemic ends. On the second proposal, normalization obtains when there is end-enforcement, or knowers fail to recognize their latitude in end-setting. The second proposal can be understood either as a distinct proposal, or as a mechanism that explains the first. In a slogan, end-enforcement normalizes by obscuring that normalization is going on.

V. CONCLUSION

To know is to normalize: we are now in a position to make sense of the most ambitious variant of the Foucaultian thesis of a link between intelligibility and normalization. This article has offered two main ways of making sense of this link. First, intelligibility can normalize by structuring salience: intelligibility-conducive concepts can make knowers’ mental lives closer to group norms in a way that is inapt given their particular ends. Second, intelligibility can normalize by targeting ends: we can set or give up our ends in a way that brings about homogeneity in an individual or collective stock of ends, or in a way that involves our latitude being obscured from us. In both cases, normalization or epistemically bad standardization follows from the dynamics of conceptualization. For the very success of concepts depends on their ability to standardize what we regard as salient, and to structure salience in ways regarding which we are not conscious. Normalization is the natural byproduct of intelligibility.

So what is to be done? One response is to simply accept these downsides as the cost of doing (epistemic) business in a world where we have finite epistemic resources. Moreover, it is consistent with this article’s argument to hold that the costs of intelligibility are always worth sustaining for the sake of moral and political benefit.

Since normalization follows from intelligibility, and intelligibility is unavoidable, it might seem that normalization is also unavoidable. This is correct, but the scope of normalization can be restricted. One way to do this is by opting for what I call a zone of disarticulation. Such a zone can be understood along individual or collective lines. For an individual, a zone of disarticulation can be an event or area of life that one refuses to subject to conceptualization, or a shorthand for a justified refusal to use some putatively apt concept. On a collective level, a zone of disarticulation corresponds to a group decision not to invest resources in the task of developing a category or taxonomy for some aspect of experience or putative harm.

Disarticulation also presents a response to the problems with intelligibility that arise when a concept is correctly applied but its salience-profile is normalizing. For some, the downsides of normalization are not substantial, and it may be considered all-things-considered better to sustain normalization when talking about, say, sexual violence or sexual identity, for the sake of solidarity with others. The important point is that opting for disarticulation need not be self-deception, repression, or false consciousness. Rather, there can be epistemic reasons for disarticulation. Sontag’s refusal to describe herself as bisexual or a lesbian, and Maria’s refusal to describe what happened to her as rape, can be understood as cases of disarticulation.

Opting for disarticulation and avoiding normalization can be motivated by multiple positive normative views. For instance, one might hold that gaining knowledge is the ultimate goal, and that this strategy is best pursued by knowers with a multiplicity of ways of knowing. Or one might hold that such multiplicity is valuable for its own sake. My own preferred strategy draws on a late interview with Foucault, in which Foucault stated that one of his goals had been to “show people that they are much freer than they feel, that people accept as truth, as evidence, some themes which have been built up at a certain moment during history, and that this so-called evidence can be criticized and destroyed.”⁸¹ Intelligibility-conducive concepts can create the impression that there is no choice but to use those very concepts, and no choice but to believe the corresponding truths. For insofar as there genuinely is a fact of the matter about whether an event is rape, or whether one is a lesbian, it can seem that defending the choice not to use those concepts is simply vindicating the subject’s right to be wrong. However, if in knowing and believing, we are also taking a stand on what matters, then for finite knowers such as ourselves, “truth” and “evidence” are never sufficient to settle the question of what we should believe. One rationale for noting latitude in our epistemic lives is the Foucaultian project of showing how we are freer than we feel. The literature on hermeneutical injustice contributes to a politics of legibility that risks obscuring this point. Scott’s account in Seeing Like a State of how legibility produces standardized political subjects serves as a warning against projects of epistemic justice that function like “state fictions” in “transform[ing] the reality they presume… to observe.”⁸² Although articulating new moral concepts can bring about moral and epistemic progress, that progress also increases the scope of normalization. The more intelligible world is a more homogeneous one. Contemporary philosophers have made compelling cases for the value of bringing the world into light; I have sought to make a case for the shadows.

ACKNOWLEDGMENTS

For helpful feedback on drafts, I would like to thank Agnes Callard, Matthew Congdon, Hasan Dindjer, Rachel Fraser, Hadleigh Frost, Mikayla Kelley, Damian Maher, Martha Nussbaum, Colton Valentine, and two anonymous reviewers and the editor of this journal. I am especially indebted to Jaime Edwards for conversation and instructive feedback across multiple drafts, as well as to Cécile Fabre for comments and crucial encouragement at this project’s initial stage.

COMPETING INTERESTS

The author declares that she has no competing interests.