The homogeneity of the “we” who make judgments deemed to be “obvious” in philosophy and elsewhere has been of great concern to feminist philosophy. … [T]hose who are non-dominantly situated are no strangers to the experience of finding obvious what others do not … it is reason to investigate whose interests are being served by some things appearing more obvious than others, to diversify the “we” of philosophers so as to expand philosophical attention to the world, and to engage in orienting knowledge work, finding ways of making what is obvious (particularly to those non-dominantly situated in the world) more obvious to others (particularly those dominantly situated in the world). (Polhaus Reference Pohlhaus2015, 11–12)
Introduction
The landscape of contemporary analytic philosophy of language is expanding, and this is a good thing. More and more analytic philosophers of language are concerned with the semantics and pragmatics of social and political terms. This is not to say that this hasn’t been done before. In fact, that this has been done before is one of the central points that motivate this paper. As Lynne Tirrell writes in Reference Tirrell, Jaggar and Young1997, “it is not news that language is an instrument of oppression.”Footnote 1 This paper aims to provide some suggestions for analytic philosophers investigating socially significant terms, working in social philosophy of language, and may be particularly relevant to those who are also committed to fostering inclusivity in philosophy through their methodologies. This is a work of metaphilosophy and methodology, whose aim is to continue and amplify the work of many others who are bridging social philosophy, feminist theory, and analytic philosophy of language.
Sentences like (1)–(3) are examples of appeals to linguistic intuition in what I call socially significant language:
(1) * I'll talk to the wop and the Italian alone. (Anderson and Lepore Reference Anderson and Lepore2013)
(2) Sue believes that bastard Kresge should be fired (#I think he's a good guy). (Bolinger Reference Bolinger2015)
(3) #Smith is an inner city pastor who is from, works, and lives, in the suburbs. (Khoo Reference Khoo2017)
The broad aim of this paper is to zero in on the use of linguistic intuitions when it comes to socially and politically significant terms (such as but not limited to: slurs, normative generics, racialized language, and code words). In section one, I give a brief background of contemporary philosophical debates around methodology and the use of philosophical intuitions—focusing specifically on linguistic intuitions. I argue in section two that social situatedness matters when it comes to linguistic intuitions about certain kinds of socially significant language. I then review the feminist epistemology literature on the importance of situatedness in accessing different types of knowledge (section three). And I argue that this applies when analytic philosophers theorize about socially significant language. I focus mostly on the discipline of analytic philosophy in this paper, but I hope these conclusions can be wide-reaching. I conclude with some methodological suggestions that involve (i) asking the philosopher of language to recognize her own situatedness, (ii) taking into consideration the linguistic intuitions of those who are in more frequent contact with the language that is being analyzed, and (iii) making sure the conclusions that we draw are appropriately constrained by the methodologies we use.
My arguments here are both descriptive and normative. The descriptive claim, which I argue for in sections two and three, is that situatedness often matters when it comes to intuitions about socially significant language. My normative claim, in sections two and four, is that theorists should pay more attention to situatedness when it comes to intuitions about socially significant language. This involves, among other things, being aware of one's own situatedness, and seeking out intuitions, theories, and accounts of language use from those whose situatedness brings them into contact with the linguistic expressions under investigation.
My final normative claim, in section four, is that if analytic philosophy as a discipline wants to stop alienating those from marginalized and underrepresented backgrounds and communities from the discipline, then analytic philosophers should avoid writing things that alienate and exclude them. When linguistic intuitions are appealed to in non-social philosophy of language, it is generally taken for granted that not everyone will agree with them. And this is not damning. If the people who disagree are enough of a minority—if the room is polled and only a few people disagree with the speaker's intuitions—then the speaker is generally licenced to continue relying on that intuition. But when it comes to intuitions about social philosophy of language, we need to pay attention to the social identities of the few who disagree. It matters if all the people in the room who share an intuition about a given slur have never been the target of that slur. And when the only two or three people who do not share that intuition have been targets of that slur, we should pay attention. Analytic philosophers must realize that we are already predisposed to ignore non-dominant intuitions. As Luvell Anderson (Reference Anderson2017, 15) tells us:
Being dubbed “the standard” [language or dialect] already confers a kind of super-legitimacy on this particular language variety, and by extension, its associated identity … Being in such a position disincentives its adherents from exhibiting the sort of humility necessary to learn from and possibly embrace alternative values.
This means, or so I will argue, that philosophers who are dominantly situated need to be extra attentive to the forms and interpretations of the socially significant language that they investigate. This is especially so if that language is used by and toward people and groups of which they are not a part.
I. Intuitions
Analytic philosophers famously mean different things by “intuition” (see Stich and Tobia Reference Stuart, Fehige and Brown2016; Cappelen Reference Cappelen2012; Nado Reference Nado2014, Reference Nado2012; Bealer Reference Bealer, DePaul and Ramsey1998). I am interested in its use as meta-philosophical term that denotes, roughly, the pre-theoretic judgments or background beliefs that philosophers rely on—or purport to rely on—to garner support for a given conclusion. Examples include the judgment that something isn't knowledge, even though it is justified true belief (Gettier Reference Gettier1963); you don't know the bank is open if your savings depend on it (DeRose Reference DeRose1992); it's permissible to divert the trolley to save five people, but not to harvest the organs of one healthy human to save five others (Foot Reference Foot1967). Philosophers use their intuitions about these cases to draw conclusions about a given philosophical theory: for example, knowledge is not solely constituted by justified true belief; knowledge is context-dependent; there is a moral difference between intentional and foreseeable killings; and so forth. Such examples pervade analytic philosophy.
My aim in this section is to illustrate the kinds of concerns raised about philosophers’ uses of linguistic intuitions, before raising some of my own concerns about intuitions regarding socially significant language. To narrow it down, I am interested in the methodology of philosophers’ and linguists’ linguistic intuitions. Linguistic intuitions are, roughly, judgments of truth, felicity, grammaticality, sense, or nonsense (Stich and Tobia Reference Stuart, Fehige and Brown2016, following Chomsky Reference Chomsky1986). Linguistic intuitions are intuitions about “linguistic, or language-like, items or their use” (Maynes and Gross Reference Maynes and Gross2013). Unlike philosophical intuitions more broadly, linguistic intuitions have the convenience of being formally demarcated for us in roughly standardized ways. Fairly long-standing tradition in linguistics allows us to categorize our linguistic intuitions about sentences with marks like ‘#’ for semantic infelicity or for unavailable or anomalous reading, ‘?’ for mixed intuitions or uncertainty, and ‘*’ for grammatical infelicity. We will call whatever is indicated by these kinds of markings linguistic intuition.
For example, in the following sentence
(4) # A student generously offered to car me home after the dinner. (Armstrong Reference Armstrong2016)
Armstrong expresses (and takes his readers to share) the linguistic intuition that sentence (4) is not semantically appropriate. He says of sentences like (4): “Even if audience members can figure out what a speaker would be trying to express … the denominal verbs occurring in these sentences are considerably more marked.”
Question marks are used to indicate a similar kind of confusion.
(5) ? John, too, will never go to Paris. (Chemla Reference Chemla2008)
Chemla uses the question mark to show that in sentence (5), it is difficult to accommodate the presupposition triggered by the word “too” without additional context. This sentence is a confusing (and infelicitous) way to indicate that John and someone else will never go to Paris.
Finally, the asterisk is used to indicate lack of grammaticality.Footnote 2
(6) * How sick were there the children? (McNally Reference McNally, Portner, Maienborn and Heusinger1997, cited in Azzouni Reference Azzouni2022)
In sentence (6), McNally illustrates that the word “there” can’t be used as an island; “there” can't be placed in that part of the sentence and continue to make grammatical sense. She notes that sentence (6) is ungrammatical and uses this observation to draw syntactical conclusions about the word “there.”
Linguistic judgments like those above are typical and commonplace in linguistics and analytic philosophy of language. The rough methodological formula seems to be: (i) authors note a feature of language they are interested in, (ii) express their intuition about this feature, with the expectation that others (will) share it, and then (iii) draw some kind of broader conclusion about language using those intuitions.Footnote 3
Here are some more features of linguistic intuition, as I am understanding them: (a) They are not exclusively held by philosophers or linguists (Antony Reference Antony and Barber2003; Maynes and Gross Reference Maynes and Gross2013). Any speaker of a language can have linguistic intuitions. Philosophers are interested in the linguistic intuitions of competent speakers of the language they are analyzing. A linguistic intuition is the kind of thing experimental philosophers of language test when they present subjects with sentences and ask them questions like “does this seem true?” “does this sentence mean X?” “does this sentence sound like a generalization?” and “if X is true, then is this sentence false?” I take it to be the case that when philosophers of language rely on their intuitions about a certain sentence or class of sentences, they rely on the implicit premise that other people share those intuitions.Footnote 4 (b) They are more or less synonymous with “we would say that” (following Cappelen Reference Cappelen2012's exegesis of Wittgenstein). For example, if I have the intuition that sentence (3) is ungrammatical, I could express it by saying “we would say that sentence (3) is ungrammatical.” (c) They are purportedly pre-theoretic (Maynes and Gross Reference Maynes and Gross2013). Linguistic intuitions are meant to capture the speaker’s “natural” or “immediate” sense about the terms in question. (d) They are a kind of judgment (Maynes and Gross Reference Maynes and Gross2013; Devitt Reference Devitt2006; among others). Despite being pre-theoretical, intuitions are still evaluative. The primary intuitions I am focusing on are judgments of sense or nonsense, grammaticality, and truth or falsity.Footnote 5
So far, I have given a rough delineation of what sorts of things linguistic intuitions are. Next, I will distinguish two ways in which linguistic intuitions are used in philosophical methodology. I take these to be two of the primary ways in which linguistic intuitions operate in contemporary analytic philosophy of language (insofar as they are the most relevant and common). I will call these (i) introspective intuitions, where the speaker reflects on her own usage of the term in question, and (ii) predictive intuitions, where the speaker does her best to predict what the intuitions of the general linguistic community will be.Footnote 6 I will show that, on either construal of linguistic intuition, extra measures of care need to be taken when an author chooses to rely on methods of intuition to theorize about socially and politically significant language in particular.
Introspective intuitions are invoked when a theorist notes her own usage of a term (or how a given sentence sounds to her ear) and subsequently draws conclusions about the term in question. That is, she uses herself as a data point from which to draw significant conclusions about the language. Maynes and Gross (Reference Maynes and Gross2013) discuss this worry in the context of linguists using themselves as subjects. As they aptly put it: “the image of a lone linguist mulling over a sentence she has mumbled to herself is not so far off the mark” (p. 722). Azzouni Reference Azzouni2022 calls this kind of methodology the narcissistic model of intuition-mongering. Here are some worries he raises:
I’d expect rather a lot of variation in people’s intuitions about what words mean … precisely because of personality type, cultural factors, perhaps gender, etc. (21)
Roughly, this is the kind of intuition use that is typified by disagreements in a philosophy talk that go: “well that sounds good to me,” and “well, that doesn't sound good to me.”Footnote 7 Recent work in experimental philosophy has found that the intuitions of individuals (philosophers and non-philosophers alike) differ to a significant level based on certain demographic features (Buckwalter and Stitch Reference Knobe and Nichols2014; Machery et al. Reference Machery, Mallon, Nichols and Stitch2012).Footnote 8 Azzouni notes that it is no surprise that intuitions differ along these individual lines—citing personal histories, idiosyncrasies of language use, and certain kinds of cultural and social factors that impact the way an individual uses and interprets language.
What’s only been shown by x-philosophical results is the (not unexpected) fact that truth-condition intuitions vary according to rather tame causal factors such as demographic factors, and (of course) the various ways our competence can be weakened by environmental factors … (Azzouni Reference Azzouni2022, 22)
Azzouni does argue, however, that philosophers’ use of linguistic intuitions can be vindicated. While we have reasons to cast doubt on the practice of using one's own linguistic intuitions as the full story, much can be gained by the kind of intuition use that accurately predicts the usage of most speakers of the language we are analyzing. Citing the distinction between relying on one's own intuition and developing a skill for predicting the intuitions of others’, Azzouni plugs for the latter as helpful methodology.Footnote 9
Azzouni calls the latter skill language-usage expertise: defining this as not only expertise about how to use the language (what most people call linguistic expertise), but also expertise with respect to how others use the language. And it is not impossible for philosophers (and others) to have this skill:
there are good reasons to think language-usage expertise is already exhibited among language users (and some philosophers) … Some—but not all—writers of fiction are very good at depicting how members of various demographic groups speak. (22 n. 31)
There is a skill to be found in predicting how others use language. This is a version of the philosophical expertise response to skeptical concerns about philosophical intuitions in general (construed more broadly than linguistic intuitions) (see Devitt Reference Devitt2011; Nado Reference Nado2012; Machery et al. Reference Machery, Mallon, Nichols and Stitch2004; among others). Proponents of the philosophical expertise view hold that philosophers, as experts with respect to philosophical domains, are licensed in prioritizing their philosophical intuitions—just as mathematicians may prioritize their mathematical intuitions. This is not to say that philosophers get to prioritize their linguistic intuitions because they are experts; rather it is to say that philosophers might be able to prioritize their linguistic intuitions because they have the capacity to be the kind of expert on others’ speech (much like successful writers of fiction and film have this capacity).
There are reasons to worry about the predictive model of linguistic intuitions. Certain philosophers may succeed in relying on the kinds of intuitions that their readers will agree to—but with no guarantee that this will match the intuitions of language users who are non-philosophers (or members of groups that are statistically underrepresented in academic analytic philosophy). They might be thinking (and intending) that their intuitions will match those of all (competent) speakers of English, when in fact it applies just to men, or North Americans, or non-marginalized speakers. I will say more about this in the next section.
Another worry about the predictive model of intuitions is that, even at its best, it may be no match for empirical and experimental work that would actually corroborate whether or not the philosopher's linguistic intuition successfully predicts or matches those of other speakers. And if this is so, accompanying worries will include those regarding the methodologies of experimental and empirical philosophy.
II. Intuitions about socially significant terms
Given the observations above, we might think that in certain philosophical settings, linguistic intuitions about the meaning and pragmatic force of certain terms will not reflect everybody's intuitions. Rather, they will reflect the intuitions about those who are using or assessing the examples at the time. I will argue that this is especially salient when dealing with socially significant language: language that particularly targets, denigrates, or stereotypes a given social group group. This is my descriptive claim. The meaning, communicative force, the harms done and conveyed, and the interpretive significance of a term “lands” differently depending on whether the hearer is targeted by such language, for instance.Footnote 10 My normative claim is that the conclusions drawn from relying on intuitions in these cases will be controversial, and philosophers should avoid relying exclusively on their own intuitions about socially significant terms, especially if they are not regular users or recipients of (or otherwise in regular contact with) those terms.
Group-sensitive language
Most of us are aware that a word like “slut” means different things to different people. For many, it is an insult. For some, it is empowering (Carr Reference Carr2013). For example, in several empirical studies, Gaucher et al. (Reference Gaucher, Hunt and Sinclair2015) found that women who read vignettes containing supportive uses of the term “slut” in SlutWalk marches were less likely to endorse rape myths. Galinsky et al. (Reference Galinsky, Wang, Whitson, Anicich, Hugenberg and Bodenhausen2013) found that self-labeling with the term “slut” encouraged self-respect. Other theorists have pointed out that empowerment is most readily accessible for white and middle/upper-class women (Cooper Reference Cooper2011; Valocchi Reference Valocchi2005), and that “slut” continues to demean and be unreclaimed for women of color.Footnote 11 For other people in other contexts, “slut” is a reclaimed term of endearment (Attwood Reference Attwood2007). In some contexts, the term is just synonymous with “promiscuous,” with no positive or negative valence.Footnote 12 This does not change the fact that most people will recognize “slut” as a primarily derogatory term. However, it does affect things like how the derogatoriness impacts the semantics (see Example 2, below).
It is not my place to teach philosophers that such a word will evoke different reactions from different people in different contexts, nor am I claiming that they do not know so. My claim is the more modest one: that individuals’ linguistic and philosophical intuitions about the word “slut” will differ depending on their social situatedness, and that matters to how we as philosophers theorize about words like “slut.” Many other terms are like this, too. Here are some examples of language whose interpretations might differ based on whose intuitions are being consulted.
Example 1: generics
Generics are expressions whose meaning and communicated content can change depending on who is speaking and who is listening. Take a normative generic discussed by Sally Haslanger (Reference Haslanger2014): “women stay at home and raise families.” A historian teaching about the nineteenth century, for example, may say this and convey a descriptive general fact about gender and division of labor at that time. A disapproving grandparent saying the same sentence to his university-bound granddaughter, on the other hand, conveys something else: something like “women should stay home and raise families,” or, “it is good and normal for women to stay home and raise families” (Leslie Reference Leslie2015; Haslanger Reference Haslanger2014). That is, depending on the context and features of the hearer and speaker, the same generic sentence will be descriptive in one case, and normative in another. The context-sensitivity of generics has been well-established (Sterken Reference Sterken2015), but here is a slightly different example about how intuitions about the generics will differ depending on who is speaking and who is listening.
In a public lecture, the following sentence pair was contrasted to show that cognitive bias influences the way people interpret generic sentences.
(7) Muslims are terrorists.
(8) # White men are terrorists.
In the lecture, (7) was judged by the speaker as felicitous, and contrasted with sentence (8) which was taken to be infelicitous (i.e., not heard as a generic claim, but rather an existential one). The context of such a discussion is that it is noteworthy that (7) is judged to be felicitous as a generic while (8) is infelicitous, despite the fact that more acts of terrorism in the United States have been committed by white men than by Muslims.Footnote 13 The speaker claimed that audiences tend to immediately hear (7) as a generalization about Muslims, while (8) has an unavailable or difficult to access reading. That is, (8) comes across as false, or incoherent.
My contention is that judgments of the felicity of (7) and the infelicity of (8) will vary greatly depending on who is judging the sentence.Footnote 14 Outside of contexts restricted to white communities, (8) is both a commonplace and felicitous normative generic statement. Without being explicit about it, the speaker was assuming a dominantly situated—in this case white, non-Muslim, English-speaking, and likely global northern—audience, and relied on linguistic intuitions that a dominantly situated audience would affirm.Footnote 15 Yet, outside such a context, it is much less likely that audiences would judge (8) to be infelicitous.Footnote 16 This matters for the accuracy of the conclusions drawn by such examples: both about the semantics and pragmatics of generics, but also about the implications for social justice. If, for example, the intention was to make claims about cognitive and implicit bias by contrasting these sentence pairs (see Wodak et al. Reference Wodak, Leslie and Rhodes2015 for a similar discussion), then a difference in the reading of (8) could make the difference between identifying an explicit or implicit bias. At the very least, acknowledging the context-sensitivity of intuitions about (8) could help narrow the speaker's conclusion to a certain community: those who do have the intuition of infelicity.
Example 2: slurs
In a public lecture, a philosopher claims that sentence (9) below entails sentences (10) and (10’).Footnote 17 The philosopher is arguing against a view of expressivism about slurring terms, using their intuition about the entailment from (9) to (10) to show that there seems to be something truth-apt about (9):
(9) There are three sluts in the building.
(10) There are three women in the building.
(10′) There are three people in the building.
Just as with the example above, I hope to show that whether or not the entailment between these sentences holds depends on who you ask. There are those who deny that sentence (9) would entail (10) or (10′) because of the dehumanizing nature of the word “slut.” This could look like the following: you have been a target of this slur. You have found it so dehumanizing that you wouldn't draw the inference from (9) to (10) or (10′). Instead, you think the person who said this says there are three sub-humans in the building. In this case, you have the intuition that the entailment does not hold because you have experienced someone using the term “slut” to slur in a dehumanizing way.Footnote 18 And there are those who would deny the intuition that (9) entails (10) because gay men are often the target of this epithet, too. In this case, rejecting the original intuition would look like this: you have experienced or observed that non-binary people and men often get called “slut,” so the entailment doesn't hold. In this case you have a different intuition. You deny the entailment from (9) to (10), but you hold that (9) entails (10′). These two cases of having different intuitions do not rely on a hearer's values or beliefs. Nor do they involve a failure of the hearer to understand or imagine that there could be specific misogynistic uses of (9) that do entail (10) or (10′). They show, instead, that having the intuition that (9) entails (10) and (10′) is not necessarily widely shared, and so does not give the whole story about the term “slut.” They also show that denial or endorsement of the intuition can differ depending on the social situatedness of the person evaluating the sentences. So, there is a divergence of intuitions about whether the entailment holds. Yet the intuition that there is an entailment is being used as data to support a conclusion about the semantics of slurring terms. As a result, the semantic conclusion only takes into account some of the data.
Example 3: derogatory terms and neutral counterparts
Certain swear words, slurs, or insults can have different meanings or communicate different things depending on the speaker, hearer, and relationship between them. The reclaimed usage of words like “slut” or “slutty,” for example, was primarily shaped by individuals who were members of the group targeted by the derogatory use of “slut” (Gaucher et al. Reference Gaucher, Hunt and Sinclair2015). So too for reclamations of words like “queer” (Clare Reference Clare1999; Valocchi Reference Valocchi2005). So-called “neutral terms” are also susceptible to meaning shifts depending on who is using them (and in what context). Some racial slurs have very different communicated contents depending on whether the user and recipient are members of the targeted group or not.Footnote 19
The following is an example of a philosopher drawing on their intuition about neutral counterparts: terms that non-pejoratively pick out the same group that a corresponding slur or derogatory term is used to target. This philosopher makes a judgment from linguistic intuition that the sentence below is ungrammatical because it is analytically false:
(11) #African Americans are not Americans.
In this example, the author assumes that sentence (11) misfires. It's either false, meaningless, or otherwise incoherent, because it is just analytically true that African Americans are Americans.
But pausing a moment to figure out whose perspective is informative, we might find that our own (hopefully) non-racist intuitions are less useful when it comes to judging derogatory language. A google search of the above phrase yields 57,400 results.Footnote 20 This should make us ask on what basis and according to whom is (11) infelicitous.Footnote 21 This matters for philosophical reasons, as well as social and political ones. The philosopher's idea here is that the term “African-American” is a neutral counterpart for other derogatory terms that are used to slur Black Americans. As their thinking goes, a variant of sentence (11) that contained a derogatory term rather than the neutral counterpart would sound more felicitous. Thus, one point of example (11) is to show that neutral counterparts don't co-pattern in a way that derogatory terms do. But, as I hope I've shown, (11) is not a hashed (#) sentence for everyone; in the context of anti-Black racism in the United States, many people use it, with no regard for a philosopher's intuition about analyticity or lack thereof.Footnote 22
Code words and dogwhistles are another set of terms whose meaning (or communicated content) varies depending on hearer or speaker. A phrase like “family values” is meant to sound innocuous enough to those who aren’t part of the target audience of it: something like valuing positive relationships with family members. But to its target audience, it communicates something like: the speaker adheres to the same kinds of right-wing Christian ideology that you do (Saul Reference Saul2017; Khoo Reference Khoo2017, Mendelberg Reference Mendelberg2001). So, a given individual's intuition about the meaning of a code word will differ based on whether or not one is in the intended or target audience, or more of a bystander (Saul Reference Saul2017).
In sum, social situatedness matters when it comes to the kind of familiarity and competence with language that partly underlies linguistic intuitions about socially significant language.Footnote 23 Briefly, here’s how to connect it to group membership, while acknowledging that group membership and situatedness are not the same thing.
First, group membership often plays a role in constructing or determining the meaning of a socially or politically significant term. Take instances of reclamation that come from members who self-identify with the extension of the term (see Jeshion Reference Jeshion2013a). Susan Stryker, writing about the history of the term “transgender,”gives us an example of this:
Transgender itself was a term then undergoing a significant shift in meaning … By the early 1990s, primarily through the influence of Leslie Feinberg's 1992 pamphlet Transgender Liberation …, transgender was beginning to refer to something else—an imagined political alliance of all possible forms of gender antinormativity. It was in this latter sense that transgender became articulated with queer. (Stryker Reference Stryker2008)Footnote 24
In cases like these, one's introspective intuitions about the meaning of the term might differ depending on whether one is a member of the impacted group or has the relevant situatedness. And one's predictive intuitions could be more or less accurate depending on how much one knows about the term's use by members of groups who come into regular contact with the term, either as a user or as a recipient. This means, as I will argue in the next two sections, that theorists should listen to people who are marginally situated.Footnote 25 Another way to put this is that better theorizing about these terms requires access and attention to diverse linguistic practices and norms. In-group membership or proximity—being a member of a group that uses the term to self-refer, or being close to such groups—can facilitate gaining explicit and implicit instructions and knowledge about how to interpret certain terms: this includes processes of immersion, interpreting in-group humor, and being proficient in slang.Footnote 26
It also means theorists should pay close attention to how people use hurtful and bigoted speech. Geoffrey Nunberg (Reference Nunberg, Fogal, Harris and Moss2018) is an excellent example of someone who looks at the speech patterns of slur-users in a way that is helpful and illustrative. This brings me to my second point, which is that membership within an antagonistic out-group can sometimes play a constructive role in determining the semantics of a socially significant term. Members of groups—often powerful—that are hostile towards members of targeted groups—often less powerful—sometimes come up with and determine the use of language that denigrates members of those targeted groups. Here, think of oppressive language, racist terminology, and code words (Butler Reference Butler1996; Moody-Adams Reference Moody-Adams1997; among others). Members of dominantly situated and non-antagonistic out-groups can also play a role in constructing the meaning of social terms (see Hacking Reference Hacking2003 on the category of woman refugee, for example). By “antagonistic out-group,” I mean people who are not members of the group and who are antagonistic towards the group. (See Kendi Reference Kendi2017 and Nunberg Reference Nunberg, Fogal, Harris and Moss2018 for arguments that slur users play a role in defining the slurring term). As many people in both of these categories—recipients and users—are neither the most-represented nor the most outspoken among analytic philosophers, the failure to pay close attention to intuitions of members of these groups can lead philosophers to draw incorrect, incomplete, or otherwise unsubstantiated conclusions about socially significant language.Footnote 27
Situatedness and intuitions
If we accept the above—that certain kinds of terms that are deeply philosophically interesting are also terms whose communicated content differs depending on who is speaking and who is listening—then there are (at least) three important questions that philosophers who engage in theorizing about such language need to ask ourselves. First, what kinds of insight and experience do we, as we are currently situated, have about such terms, such that our linguistic intuitions about them are instructive and informative? Second, what kinds of insight into and experience of these terms do others have? Third, are we consulting those people and their intuitions and reading and citing their work on these topics? The answers to these questions matter when it comes to the methodologies we endorse and the scope of the conclusions we draw.
Being a member of a marginalized or oppressed group generally involves having certain lived experiences that are not had by members of dominant groups. These experiences will not all be the same, but they will more often than not be more informative than the experiences that members of the dominant group have had. For example, the experience of racism, or sexism, or classism, will manifest differently for different people, especially along different intersections of oppression, but they will, for the most part, be notably absent among people who don't face these particular kinds of oppression and discrimination (Collins Reference Collins2015). And so too for linguistic manifestations of this kind of oppression. For example, only some people have been recipients of slurs. That kind of experience matters for our linguistic and philosophical intuitions, and informs them.Footnote 28 What I hope is that theorists keep such complexities in mind when engaging with their own and others’ intuitions about socially significant language. Again, being a member of a marginalized or oppressed group is neither sufficient nor necessary for having the relevant intuitions that should be prioritized, but it often is informative. There are other things to consider, including: interests, background, and diversity of identities and experiences within any given group.Footnote 29
The way this should factor into philosophy of language is manifold. I take all the historical suggestions from Labov (Reference Labov1972, Reference Labov and Austerlitz1975), and more recently from Hudley et al.'s (Reference Hudley, Charity and Bucholtz2020) work on racial justice in linguistics. For example, from Hudley et al. (Reference Hudley, Charity and Bucholtz2020, 200): “To be adequate, a linguistic theory of race must incorporate the perspectives of linguistic researchers of different methodological approaches and racial backgrounds and must also draw on theories of race in neighboring fields…”. Consulting as many people as possible while avoiding tokenizing is an important first step. The second suggestion is to be explicit about where the theorizer themself stands in relation to what they are theorizing: in terms of lived experiences, interests, and to whatever extent possible, being upfront about our own possible biases and limitations. I discuss this further in following sections.
So far, I've given reasons to pay more attention to the linguistic intuitions of people who are situated in specific ways. A stronger thesis would hold that we also have reason to pay less attention to the intuitions of those who are dominantly situated. One way to argue for the stronger claim would be to observe the pitfalls of theorizing about language when one is not a member of the group affected by such language (and when one does not consult members of those groups). As Nancy Bauer (Reference Bauer2015, 105) points out: “In failing to attend carefully to how real people actually speak or what phenomena in the world (pornography, say) are actually like, what we [philosophers] say is, at worst, wrong and, at best, hollow.” Anderson and Lepore (Reference Anderson and Lepore2013, 7–8) make a similar point:
In academic discussions and in the quiet of a study, it's easy to convince oneself (we confess on occasions we have) that particular uses of slurs are inoffensive. We couldn't have written this paper had we not. As a safeguard against such inurement, we strongly urge you always to ask yourself how a targeted member, perhaps accidentally overhearing you, would react to your usage. You'll find, as we have, that much of what seems suitable is definitely not.
This advice from Anderson and Lepore is not only prudentially and morally important, but illustrates a further positive upshot: it's not that certain things are completely inaccessible to dominantly situated theorists—or to theorists who are not users or targets of the language in question—but that it takes more work than a priori theorizing for a dominantly situated theorist to internalize what certain experiences are like for others. I follow those feminist scholars who argue that “differences in social location might make some things appear more obvious to [marginally situated individuals], but that these things can be made obvious to [dominantly situated individuals], and they ought to be made more obvious to all philosophers” (Pohlhaus Reference Pohlhaus2015, 15).Footnote 30
Bauer, Anderson and Lepore, and Pohlhaus give us another important normative takeaway. It is not that dominantly situated philosophers and theorists cannot theorize responsibly about socially significant terms and language. It's that, in most cases, it requires more than relying on one's default judgments. I do not claim here that the linguistic intuitions of dominantly situated analytic philosophers should be dismissed or devalued, but I do maintain that the linguistic intuitions of dominantly situated analytic philosophers are not going to tell us the whole story. It will take further reflection and will require methodology other than introspecting on intuitions to gather information and arrive at conclusions that will be relevant to the questions at hand.
III. Situatedness
Why is reliance on intuitions particularly problematic for social philosophy of language? In short: it's not. Situatedness impacts our intuitions about many kinds of language. There are two reasons I focus specifically on socially significant language. First, the stakes are higher, as compared to language that does not target, denigrate, or stereotype. While this paper has focused on explicitly marked derogatory and social language, I take this point to apply to implicit discourse as well.Footnote 31 When it comes to the study of language that mostly focuses on understanding oppression and social justice, it is doubly damning if we get this wrong. First, we are at risk of getting our theories wrong. Second, there are moral considerations around making sure those who are impacted by theorizing are centered in it. The observation about language in general has been made in other ways throughout the history of linguistics, sociolinguistics, and philosophy of language. I will discuss these in this section. The rest of this paper will focus on how social situatedness influences how we use and understand language. One way to do this is to first turn our attention to ways in which social situatedness influences how we know.
Situatedness and knowing
Feminist epistemology tells us that the way one is socially positioned makes a difference to how one engages with the world, and what they know about the world as a result. Furthermore, our social position matters to how we are situated as knowers; and it influences the kinds of input, information, and evidence we get, how others treat us, and the degrees to which we are able to abstract away from our individual perspectives (Haraway Reference Harding1991; Harding Reference Harding and Harding1987; Collins Reference Collins2015; Pohlhaus Reference Pohlhaus2002; Kukla and Ruetsche Reference Kukla and Ruetsche2002). One baseline commitment of these thinkers is that situatedness positions individuals differently as knowers. Certain individuals will know about certain things because of their familiarity with them as a result of their position in the world. For example, as a general trend, tall people will know more about the phenomenology of bumping their foreheads on doorways, people living in colder climates will know more about how to walk on icy sidewalks. Analogously, individuals on food stamps will know which stores accept them and which don't; people who need gender-inclusive bathrooms will know which buildings have them and which don't; survivors of sexual violence will know the tell-tale signs and patterns of harassment and abuse.
Others (Harding Reference Harding1991; McKinnon Reference McKinnon2015; Pohlhaus Reference Pohlhaus2002; Collins Reference Collins2015; DuBois Reference DuBois1903; Anderson Reference Anderson1995, Reference Anderson2015) argue that situatedness better positions certain—marginalized—individuals as knowers.Footnote 32 That is, those individuals on food stamps won't just know which stores accept food stamps, but they will be better positioned to know this than individuals who are not on food stamps. This might seem trivially true, but consider cases where a non-marginalized person should know (or notice) what the marginalized person knows, but doesn't because their privilege obscures access. For example, most people working in a given building are in a position to know whether that building has gender-inclusive bathrooms; but a person who needs to use them is better situated to notice whether and where there is one.Footnote 33 Similarly, survivors of sexual violence are better positioned than others to recognize instances of harassment and violence as such. This is a commitment of standpoint epistemology, which holds that:
social location systematically shapes and limits what we know, including tacit, experiential knowledge as well as explicit understanding, what we take knowledge to be as well as specific epistemic content. What counts as “social location” is structurally defined … by [individuals’] location in a hierarchically structured system of power relations: by the material conditions of their lives, by the relations of production and reproduction that structure their social interactions, and by the conceptual resources they have to represent and interpret these relations. (Wylie Reference Wylie, Figueroa and Harding2003, 31)
Following Wylie, I understand social situatedness as resulting from the interplay of social and material factors. There are different kinds of social locations and different kinds of situatedness. Some features of situatedness are determined by race, gender, class, ability, geography, and other considerations like interest, background information, or knowledge about a certain topic.
Feminists have also pointed out that interests also influence our intuitive judgments.Footnote 34 To elaborate on the above example: a nonbinary person will have more knowledge about the location (or existence) of gender-neutral bathrooms not just because of their situatedness, but also because they might have an interest in gender equity. And this is not restricted to situatedness: a cisgender person who is interested in promoting the interests of transgender and gender-nonconforming people will have more of a motivation to seek out this knowledge.Footnote 35 Relatedly, a survivor of sexual violence might not just know what to look for in identifying cases of violence because of their experience, but they might also have an active interest in preventing further abuses. And this interest plays a role in the epistemic practices of the survivor. In this way, interests are an important part of situatedness and its effects on knowing.
Situatedness and linguistic intuitions
There are two ways to tie linguistic intuitions into the broader discussion of standpoint epistemology and situatedness. One way to do this is to frame an individual's linguistic intuitions in terms of some epistemic feature of language. This could be knowledge of linguistic practices, norms, meanings, etc. It could also be access to evidence. Going this route would mean treating intuitions as a kind of epistemic phenomenon. A related way to go is to extend the analogy more broadly: situatedness matters to the way in which one theorizes about X, so that theories of X reflect contingent and non-universal features of the embodiment of the theorists who espouse it (Antony Reference Antony and Moser2002). And since relying on linguistic intuitions is part of theorizing about language, intuitions are implicated in our epistemic practices of theorizing.Footnote 36
Situatedness matters when it comes to socially significant language. But it's also the case with language more generally: think of dialects, slang, words meaning different things in different parts of one country with the same official language (like “pop” and “soda”), and syntactical differences across different sub-communities of speakers of English. All of this matters when we use linguistic intuitions as evidence.
One way this matters is with respect to judgments of linguistic felicity. The sentence:
(12) Now I eat it with a spoon anymore.
will sound grammatical to someone from Pennsylvania or Kentucky, but not to someone from Rhode Island or Washington state. Footnote 37 That is, some people have the linguistic intuition that (12) is grammatical, and some do not.Footnote 38 So, when we draw conclusions about the semantics and syntax of “anymore,” the conclusions will be constrained by the linguistic communities whose judgments we are considering. If linguists did not consider (or discounted) the linguistic intuitions of speakers from rural Pennsylvania and Kentucky, they would not have the full picture about positive “anymore.” This first way of linking up situatedness to intuitions involves different individuals using and interpreting the same expression in different ways.
There's a second way of linking up situatedness to linguistic intuitions: who is speaking—and to whom—matters when we are determining the communicated content of what is said.Footnote 39 Sally McConnell-Ginet, following Paul Grice, tells us that “meaning depends not just on the speaker but on a kind of relation between the speaker and the hearer” (McConnell-Ginet Reference McConnell-Ginet and Cameron1998, 200). Jennifer Hornsby, in a similar spirit, proposes a supplement to traditional semantics that gives an account of “saying something to someone” (Hornsby Reference Hornsby, Fricker and Hornsby2000, 2, original emphasis). Hornsby criticizes the analytic tradition of analyzing meanings and sentences in a vacuum. In doing so, she argues, analytic philosophy of language underemphasizes the relationship between speaker and hearer. In particular, it underemphasizes the context-dependency of meaning on the speaker, the hearer, and the relationship between them when the expressions in question aren't already standard context-sensitive language. Here is one way she puts it, contrasting the traditions of feminist theory and analytic philosophy of language:
[In] feminist work, the use of language is treated always in a social context, in which the presence of gendered beings is taken for granted. In philosophy of language, by contrast, when modality, say, or relative identity, or reference is the topic, the subject matter is apparently far removed from any social setting … focus on semantic theories has actually helped to sustain the appearance of a gulf between philosophical treatment of language and the treatment of social phenomena. (Hornsby, Reference Hornsby, Fricker and Hornsby2000, 4)
Lynne Tirrell agrees: “The interweaving of philosophy with linguistics and literary theory makes feminist philosophy of language significantly different from traditional philosophy of language, although they share some methods and concerns” (Tirrell Reference Tirrell, Jaggar and Young1997, 140).
I think these are good diagnoses of the differences between the two fields and their methodologies. And one way to put my concern is that, increasingly, certain philosophy of language that investigates social terms still does so in the paradigm of philosophy of language that is “far removed from any social setting” (Hornsby Reference Hornsby, Fricker and Hornsby2000). This is problematic because, among other things, such language is deeply enmeshed in the social world.Footnote 40 The practice of analyzing language removed from social context can be a valuable part of the analytic tradition. But it shouldn't be with this kind of subject matter. Hornsby refers to a gulf between analytic philosophy and feminist theory. But there needn't be one. Philosophy of language has much to gain and much to contribute by engaging with the kinds of social phenomena that appear to be highly linguistic (such as slurs, generics, code words, fig leaves, slang, insults, epithets, etc.).
Linguistic communities
In the previous section, I used positive “anymore” as an example of how linguistic intuitions of felicity can differ depending on the speaker's background (specifically, regional or geographic background). Linguists who study positive “anymore” invoke a notion of linguistic community.Footnote 41 The rough idea is that certain terms will be used differently depending on their speakers, or groups of speakers. This can be extended to syntax, pragmatics, and semantics. Hornsby gives us a way of thinking about linguistic communities and semantic theory:
if the idea of a semantic theory is to cast light on the general concept of linguistic meaning, then something general has to be said about the relations between languages (thought of now as the objects of semantic theories) and groups of speakers. We might say that a semantic theory for a language is correct only if it belongs inside an overall account of the lives and minds of the people who use the language … (Hornsby Reference Hornsby, Fricker and Hornsby2000, 4–5; my emphasis)
The point here is that language use and language meaning differ from community to community, and sometimes from speaker to speaker. Certain individuals (qua group members but also qua individuals with a certain situatedness) have more access and familiarity with certain terms. This access and familiarity can be gained by virtue of their lived experiences as members of certain (and myriad) social groups, familiarity with those groups, or social and political commitments or interests.Footnote 42
Philosophical and linguistic attunement to situatedness
Linguists and analytic philosophers of language are already methodologically attuned to situatedness in many circumstances. So, my suggestion that we pay attention to social situatedness should not be viewed as a radical divergence from the standard methodologies of analytic philosophy and linguistics. Rather, it's on a continuum with the kinds of situatedness that are standardly taken into account. Here are some other examples, hopefully familiar, where it's standard to take a speaker, hearer, or evaluator's situatedness into account.
On some leading views of the semantics of epistemic modals—words like might as in “I checked the forecast and it might be raining” and should as in “the coffee should still be hot since it was just brewed”—the truth-conditions of sentences containing these terms depends on the speaker's or hearer's evidence or knowledge base (see Fintel and Gillies Reference Fintel and Gillies2011). When I say “it might be raining,” according to these views, I am saying that it is consistent with my evidence—or with the things I know—that it is raining. So, what I know matters to the truth and meaning of this sentence. If you are in a windowless room and I am outside in the hot sun, our intuitions about the truth of “it might be raining” will differ.
Indexicality is now taken to be a standard feature of language (following Kaplan Reference Kaplan, Almog, Perry and Wettstein1989). Indexical terms like “I,” “you,” “this,” and “there” need to be interpreted within a given context in order to have a coherent meaning. Contextual parameters are usually taken to encompass at least the speaker, time, and world, but often include more (Lewis Reference Lewis, Kanger and Öhman1980; Kaplan Reference Kaplan, Almog, Perry and Wettstein1989; Mount Reference Mount, Russell and Fara2012). So here too, the situatedness of the speaker or the evaluator of the sentence matters. If I am evaluating the sentence “we've had a mild winter” in Ann Arbor in 2016, my judgment will differ from someone who is evaluating the sentence in Boston in 2015.Footnote 43
Although these are examples of different kinds of situatedness from the discussions of social situatedness and oppositional perspectives, the point here is just to show that analytic philosophy of language and linguistics can account for individual differences that bear on semantic, pragmatic, and other judgments about terms.Footnote 44
IV. Methodological upshots
The methodological upshots and advice for analytic philosophers examining socially significant language are the following: (1) read and reference non-philosophical literature that addresses the same kinds of questions about socially significant language, especially from disciplines with a history of doing so; (2) read and reference scholars and writers in communities affected by such language; (3) reflect critically on our introspective and predictive linguistic intuitions; (4) adopt methodologies from other disciplines that center these issues and inquiries, such as being forthcoming about our own situatedness as scholars; (5) perform empirical experimental philosophy and corpus work to bolster or accompany our use of predictive intuitions. Footnote 45 I will go through these in turn in greater detail.
For (1) and (2), I am imagining engagement with memoirs, texts from sociology, socio-linguistics, political theory, critical race theory, feminist and queer theory, and other disciplines, depending on the terms under investigation. For example, poet and activist Eli Clare's (Reference Clare1999) book Exile and Pride is a first-hand reflection on terms like “queer,” “crip,” and “freak”’—socially significant language used against and by (in some reclamation contexts) people in the queer and disability communities—from the perspective of someone who is trans and living with cerebral palsy. The extensive discussion of these terms and the contested reclamation of some of them within LGBTQIA+ and disability communities, and the first-hand account of lived experiences both of being slurred by these terms and of using them in empowering and reclamation contexts, would be useful to theorists working on the semantics and pragmatics of these terms. And it would provide valuable and nuanced perspective about the terms from a perspective not often cited or referenced in analytic philosophy. Investigations of socially significant language are inherently interdisciplinary, so engagement across disciplines can only make this research more accurate and wide-reaching.
For (3) and (4), I will include the important caveat that concerns about a theorist's privacy and safety should come first.Footnote 46 Especially when someone is already in a marginalized and less powerful position, professionally and in the world, it can be increasingly risky to out oneself to the world. So I will say, within limits and wherever possible, it would be good to situate oneself with regard to one's intuitions and theorizing, especially for people writing from more powerful and privileged positions. For example, Lisa Guenther's (Reference Guenther2021) excellent paper about women's prisons in Canada includes an extensive note of her experiences volunteering in correctional facilities, but also an acknowledgment of her lack of lived experience as an inmate. I hope that the advice to interrogate one's own situatedness and positionality with respect to their research can be helpful even if that situatedness not made public.
Suggestion (5) is an invitation to a more pluralistic way of doing philosophy of language. It's true that the majority of this paper has focused on intuitions, and that (5) is a departure from that. But in doing corpus work and experimental work, one is able to observe and consult how users use the language under investigation in a way that would be similar to consulting the intuitions of those who use the language.
If one of the important questions we're seeking to answer when we do social and political philosophy of language is “what is the social or political role of a given expression?” then our answer must extend beyond the answer to the question: “what is the social or political role of this expression for dominantly situated analytic philosophers?” Conflating the two questions makes our theories incomplete and further marginalizes non-dominantly situated philosophers by treating their judgments as irrelevant. So it is both in service of accuracy and of inclusivity that I make these suggestions.
A second kind of question that analytic philosophers of language ask about socially significant terms is about their meaning. Here, too, members of the discipline would be well-served to consult and incorporate the intuitions and scholarship of people who are socially situated in particular ways with respect to these terms. I have discussed this in more detail above, but concrete examples include: deferring to and otherwise centering the intuitions of communities and individuals within them who have played roles in constructing the meanings and uses of certain terms (like the reclamation of “queer,” or the construction of “transgender”), and individuals who have been on the receiving end of pejorative uses, as well as investigating, through ethnography, corpora, or first-personal accounts, those who use them pejoratively.
I suggest the dominantly situated theorist can better answer both of these kinds of questions by recognizing their own situatedness and how this affects their relationship to language. This means taking their own linguistic intuitions with a grain of salt, especially if they are not familiar, in a lived sense, with this kind of language.Footnote 47 This also means realizing that other individuals may have different intuitions about the same term.
Finally, the dominantly situated analytic philosopher should take into account the linguistic intuitions of others who are more directly impacted by and more in contact with the terms under consideration. This means engaging with scholarship and literature written by members of those groups. This could also involve careful experimental work.Footnote 48
If we are hoping to make an academic discipline more inclusive, we need to think more carefully about our methodologies. It's not just a matter of increasing bodies and representation. It's a matter of respecting and engaging seriously with non-dominantly situated work, viewpoints, and intuitions.
Acknowledgments
For invaluable feedback and conversations on this and earlier versions of the paper, thanks to: Derek Anderson, Luvell Anderson, Emma Atherton, Lauren Ashwell, Jody Azzouni, David Balcarras, Liam Kofi Bright, Carolina Flores, Sally Haslanger, Cassie Herbert, Kat Hintikka, Jonathan Jenkins Ichikawa, Alice Kelley, Zoë Johnson King, Justin Khoo, Meena Krishnamurthy, Ishani Maitra, Rachel McKinney, Deborah Mühlebach, Kevin Richardson, Haley Schilling, Jason Stanley, DeeAnn Spicer, Bob Stalnaker, Eric Swanson, Olúfẹ́mi O. Táíwò, Philip Yaure; and audiences at MIT, the Eastern APA, the New York Philosophy of Language Workshop, and the California Roundtable on Race and Philosophy. Special thanks to the anonymous referees and editors at Hypatia, and to Briana Toole for their extensive feedback and encouragement.
Samia Hesni is Assistant Professor of philosophy and women's, gender, and sexuality studies at Boston University. Their main research interests are in feminist philosophy, social/political philosophy, and philosophy of language.