Hornsby, Hom, and the Semantics of Slurs (Pt. 1)

What do slurs mean?

Some notes and important context:

I wrote the following post as a short paper for a philosophy of language class. It’s about Jennifer Hornsby’s paper “Meaning and uselessness: how to think about derogatory words.” Slurs have, in the last 30 or so years, become a subject of fascination for many philosophers of language. These philosophers hope to find a theory of the meaning of slurs which can account for their confusing characteristics (e.g., their offensive content, uses, appropriation, grammatical behavior, implicature, truth-conditions, and so on). The dialectical context in certain respects resembles that of meta-ethics in that we see a cleavage between cognitivist and non-cognitivist, expressivist and representationalist, and reductive and non-reductive views. But we also see a large cleavage between accounts that try to locate the offensive content of slurs in their pragmatics and those that locate this content in their literal meaning or semantics.

Hornsby is less concerned with providing an account than she is with showing how an apparent constraint on our theorizing reconcilable with a broadly representationalist approach to the meanings of slurs. After explaining this constraint and critiquing a handful of extant representationalist accounts, she gives a short, speech-act theoretical metaphor explaining how the representationalist can instead think about slurs. Chris Hom, in his “the Semantics of Racial Epithets” provides several arguments against the applicability and theoretical fruitfulness of Hornsby’s metaphor.

The goal of my paper was, more or less, to set up Hornsby’s project and metaphor, and show why Hom’s objections, more than anything else, impose constraints on the project of fleshing Hornsby’s metaphor out into a theory. I’m currently working on a second part of this post (coming out within the next decade or so) which will offer a sort of spiritual successor view to Hornsby’s that combines her best insights with those of Elisabeth Camp, and the philosophies of Iris Murdoch, Stanley Cavell, and John McDowell.

This paper was rushed so is pretty rough around the edges, but I thought it was still worth uploading. Hope you guys enjoy it!

Hornsby, Hom, and the Semantics of Slurs

In the early 2000s, Jennifer Hornsby developed a novel, heterodox approach for providing the meanings of slurs. Several features distinguish her approach from its rivals: modesty, skepticism, and moral qualms about the use of truth-conditional semantics in providing a theory of the meaning of slurs; the use of speech act-theoretical resources; and a disavowal of non-cognitivism. But her approach and the paper in which it is outlined are rather obscure.

In this post, I hope to clarify the obscurities of her view before considering how it fares against criticisms from Chris Hom. I conclude that not all of Hom’s attacks are serious threats to Hornsby, but that nonetheless there is much work for her to do.

Hornsby and the Ethics of Modesty

Let us briefly reconstruct Hornsby’s 2001 paper. Hornsby’s primary concern is the way a representationalist theorist of meaning can and cannot approach the meanings of slurs. In particular, conjunctivist non-cognitivism and inferentialism are shown to be poor approaches. Her responses to these views are bookended by discussion of what I take to be the meat of the paper: we ought to be modest about the prospect of offering truth-conditions for slurs in a semantic theory. This is detrimental, or so she argues, only for proponents of a pathological, overly ambitious conception of the role of truth-conditions in semantic theory.

The notions of modesty and ambition are drawn from, but do not quite follow, a similar distinction made by Michael Dummett and John McDowell in their work concerning the significance of semantic theorizing. In this literature, ambition, or “full-bloodedness,” is roughly Dummett’s position that a semantic theory for a language cannot rely on content-determining uses of concepts expressible in the object language. Or, relatedly, that our semantic theory must not presuppose an understanding of the concepts for which it serves as a theory of meaning. Someone must be able to derive an understanding of the object language via their understanding of a semantic theory or, at the very least, their understanding of the object language must consist in their competency in the practical abilities described in a semantic theory. Semantic theory must be, as it were, “‘as from outside’ content and concepts” (McDowell 2001 p. 92). Modesty is, as the name suggests, a more modest construal of the function of a semantic theory. A semantic theory, from this angle, must just specify the content of sentences or utterances to an intended audience. We can (and must, for McDowell) do so from within content and concepts.

Hornsby’s distinction is more focused on the potency of truth-conditions in a semantic theory. An ambitious view here is a view wherein an understanding of the language just consists in, or is derivable from, an understanding of the truth conditions of its sentences. One understands p if and only if one understands p’s truth conditions. The modest alternative is that truth-conditions merely state a fact about what is asserted when a sentence of the appropriate type is asserted. The value of truth-conditions is their disquotational function—they map sentences of a language to, as Hornsby puts it, “something understood by those who understand [them]” (p. 11). A modest theorist may well think that a meaning of such a sentence has been given when its truth-conditions have been given (and in fact she takes a commitment to this view to be constitutive of representationalism (Hornsby 2001 p. 4, footnote 3)), but these conditions do not exhaust or constitute what one understands when one understands the meaning of the sentence. Truth-conditions merely record the content of the assertoric use of a sentence. An important consequence of this condition for Hornsby is that an ambitious theorist, but not a modest theorist, must conclude that the exclusion of the truth-conditions of some sentence from a semantic theory implies that the sentence is meaningless.

This point matters to Hornsby because she holds that a slur does not belong in a semantic theory. This is because the truth-conditions for a sentence which uses the slur must themselves use the slur, but slurs are useless. Her reasoning here is rather simple, although it rests upon a sophisticated speech act-theoretical view of meaning. Slurs are used to bully, degrade, vilify, threaten, mock, and so on, and these uses have unavoidably colored, tinged, tainted, and rubbed off onto slurs, leaving them with an abhorrent connotative or overtonal residue. So for those people who do not wish to perform any of those acts, there is nothing they can wish to say which they can say with the use of a slur. And so, for those people, slurs are useless.

Consider the semantic theorist offering truth-conditions for an assertoric use of a slur. “‘John is a n*’ if and only if,” they continue, “John is a n*.” The latter clause is embedded in a conditional, but the derogatory content of the slur seems to project up to the theorist despite this. Compare this to a sentence from Laserohn 2007:

1. I consider John a saint. But if he ever screws me over, I’ll crush the bastard like a bug.

The speaker of (1) has not threatened or derogated John. The offensive, expressive content of the second sentence is blocked by its being conditional on John screwing the speaker over. John may even reasonably assent to (1) in certain circumstances. But, Hornsby seems to argue, slurs do not function in this way. (This should be immediately apparent if one switches out ‘bastard’ in (1) for ‘n*’.) The semantic theorist is no less subject to constraints regarding the use of derogatory language than anyone else, and so “we… cannot assent” to their use of the word in a semantic theory. On the modest view, the theorist is (as an analogy to McDowell and Dummett’s use of ‘inside’/’outside’), in a sense, theorizing within the object language and the social practices which constitute competency with that language. And so they should feel no pressure to provide a T-sentence for an assertoric use of a slur. But if we are immodest about the role of truth-conditions in our theory, then we must give slurs’ truth-conditions if they are to have meaning, and thereby use them in such a way that their derogatory content projects up to us. In trying to describe the meaning of the slur, a theorist may come too close to it; touch it with bare skin.

Hornsby gives us an analogy with which we can think about this penetrating derogatory content of slurs:

It is as if someone who used, say, the word ‘n*’ had made a particular gesture while uttering the word’s neutral counterpart. An aspect of the word’s meaning is to be thought of as if it were communicated by means of this (posited) gesture. The gesture is made, ineludibly, in the course of speaking, and is thus to be explicated, as the socially significant thing it is, in illocutionary terms. (p. 12.)

This point is difficult, but given Hornsby’s earlier discussion of uses “rubbing off” onto words, I think we can understand it as follows. Slurs are words of derogation. They are words of bullying, subordination, hatred, harassment; words for bullies, chauvinists, bigots, and harassers. If this is what they are, then we cannot grasp their meanings in isolation from this character. Hornsby explicitly accepts that a slur gives itself to these uses because of its meaning, but it seems she proposes that, to some extent, the relationship goes the other way: a slur partially means what it does. Compare this with another gesture. It is the offensive connotation of the middle finger which makes it apt for deployment in situations where one hopes to offend or provoke, and yet it is the use of the middle finger for this purpose which imbues it with this significance. The middle finger partially means an act of offense–one cannot be said to understand the gesture unless they understand the acts which it performs.

A potential successor to Hornsby’s account could be Chang Liu’s view of slurs as illocutionary force indicators (Liu 2020). Consider the following example. The phrase “John will arrive shortly” can be both confessed and promised. Confessed when, say, one invited John when they were not supposed to, and promised when John’s friend is covering for John, who is late. The phrase ‘I promise’ as it appears in the sentence “I promise that John will arrive shortly” serves as an indicator of which illocution one is engaged in. Similarly, Liu supposes, that a sentence containing a slur and a sentence containing the slur’s neutral counterpart have similar content, but the slur indicates that one is engaged in the illocution of derogation. This seems rather consonant with Hornsby’s view: a slur’s contribution to truth-conditions may be identical to the contributions of a neutral counterpart, but part of its meaning is its indication of a particular (family of) illocutionary act(s).

Hornsby’s ethical considerations have led her towards a theoretical perspective–towards views about the truth-conditions, meanings, and force of slurs. But it is worth pointing out a direct normative implication of this perspective. It is not through a (pure) use of a slur that we can counter the derogatory content of a slur, for when we use, we gesture. That is not to say we cannot utter a slur in countering a slur (e.g., “John is not a n*” or “there are no n*s”), but these uses must, in context, exhibit a sort of metalinguistic quality. It is not enough to deprive your slur of its extension, I must deprive us, our interaction, of your slur. I must object to the use of the slur–to the thought that such an act could (ever) be appropriate. Such utterances can be faithfully paraphrased so that a slur is only mentioned, and this is not so with a derogatory use. It is characteristic of Hornsby’s view that its ethical and theoretical dimensions are mutually reinforcing. Ethics and theoretical semantics are interpenetrative to Hornsby; pervade each other at every step.

Hornsby v. Hom

I want to turn now to a series of objections levelled against Hornsby. I hope, in doing so, that we can construct a list of weaknesses that a better account must patch over. Most come from Chris Hom’s paper “the Semantics of Racial Epithets,” and one objection is my own.

The first objection I want to consider from Hom is his question of how slurs contribute to the truth-values of sentences. He writes that, “If these words make no contributions, then we are faced with the unenviable consequences of truth-value gaps and having to alter the traditional rules of logical deduction” (p. 418). There is good textual evidence to think Hornsby takes slurs to have the same contribution to truth-conditions as their neutral counterparts. First is her stance, quoted above, that in using a slur, it is as though one has used its neutral counterpart alongside a gesture. Second, her account is intended for use or consideration by the representationalist. She wants to show how the features that lead one to suspect that non-cognitivism is the best approach can be accommodated in a cognitivist treatment. When she proposes the exclusion of slurs from theories of meaning, the point is not that the T-sentence for an assertion containing a slur is incoherent or that we must alter our views on truth and deduction to make it coherent. Her view is that such a T-sentence would not belong in our theory of meaning.

I imagine Hom would respond by redirecting us to the following argument he levels against pragmatic accounts: if slurs are truth-conditionally equivalent to (that is to say, coextensive with) their neutral counterparts, then

2. John is an African-American.

would entail

3. John is a n*.

Further,

4. African-Americans are African-Americans

is trivially, analytically, a priori true, while

5. African-Americans are n*s.

appears non-trivial and false.

I think Hornsby has a way out. Here is what I would like to say on her behalf: if an assertoric use of the slur constitutes an act of derogation (or indicates that an utterance is such an act), then (3) and (5) cannot be used to state a fact. That is, the presence of the derogatory content prohibits the sentence from being part of an act of reporting or stating. When we ask whether (4) is true and then ask whether (5) is true, we are, in both cases, asking whether one assents to the sentences. But what this assent consists in is not identical. Assenting to (4) may be assenting to a proposition, but (5), in a way, contains more information than (4). Hornsby would say that the inclusion of the slur in (5) limits the possible deployments of the sentence to derogatory speech acts, and so assent to (5) can never be merely assent to a proposition, but complicity in–perhaps endorsement of–an act of derogation. I suspect that this is why we generally have the intuition that (5) cannot be true.

Here, Stanley Cavell’s concept of acknowledgement finds an unexpected application. Cavell introduces the idea in his paper in “Knowing and Acknowledging” when discussing sentences like “I am in pain.” He writes (rather beautifully):

…your suffering makes a claim upon me. It is not enough that I know (am certain) that you suffer–I must do or reveal something (whatever can be done). In a word, I must acknowledge it, otherwise I do not know what “(your or his) being in pain” means. Is. … I do not mean that we also in fact have sympathy…. The claim of suffering may go unanswered. … The point, however, is that the concept of acknowledgement is evidenced equally by its failure as by its success. It is not a description of a given response but a category in terms of which a given response is evaluated. A “failure to know” might just mean a piece of ignorance, an absence of something, a blank. A “failure to acknowledge” is the presence of something, a confusion, an indifference, a callousness, an exhaustion, a coldness. (Cavell 1976, p. 264).

Pain occupies a particular place in human life. Its position is such that there is no such thing as merely knowing that it is present. A man in pain must express it–he is, in Cavell’s words, impaled upon it–and if he does not, then he suppresses it; deems it unworthy to be expressed at that moment. In either case, his response to his knowledge of his pain is necessarily expressive. A mode of evaluation is opened which is not opened for his knowledge that Hespherus is Phosphorus or that that bird is a goldfinch. That knowledge can be latent, can be trivia. Similarly, his statement, “I am in pain,” cannot be merely recorded by me. My knowledge of his pain admits of the same mode of evaluation as does his knowledge of his pain.

Perhaps, then, slurs occupy an analogous place in our discourse. (4) can simply be regarded and evaluated as a logical truth. (5) (by Hornsby’s light) introduces a type of content that demands acknowledgement. If I say nothing, I do something; let something slide. My assent or dissent or silence towards (5) acknowledges that offensive content; expresses my approbation or disapprobation, my sensitivity or my callousness, my complicity or my protest. This is why we do not wish to call (5) true—because it is not the sort of sentence that can be just stated. It is not the sort of sentence we would ever ordinarily evaluate as true or false. Our evaluation must take a wider scope than our evaluation of (4), or perhaps a different scope altogether.

Further, to say the sentences (3) or (5) are false is to say their speaker has misapplied the slur. Do we have that intuition? Hom takes it for granted that most people would answer in the affirmative, but I am less certain. Slurs do often seem to get something representationally correct. Suppose a man bets that “the next person this company hires will be a n*.” We know immediately under what conditions he will have won the bet (Camp 2018, p. 32). He wins just in case the next person the company hires is black. And, if John is black, then (3) seems like a linguistically competent use of the slur in that it targets the right group. Put differently: our evaluation of (3) if John is black is not the same as our evaluation of (3) if John is white. In the latter case, a speaker has used the slur incompetently.

There is a kernel of truth to Hom’s intuition, however. The speaker of (3) does also seem to have gotten something cognitively wrong. Hornsby can partially accommodate this by just pointing out that some level of cognitive error must be in place for the racist who uses the slur—a distorted view of a social group, incorrect ethical beliefs. But, to Hom (and Camp’s) point, Hornsby’s view renders the connection between such cognitive shortcomings and the use of a slur rather accidental. The use of a slur may suggest epistemic failure, but it does not quite express it. Given the systematic correlation we observe between acts of derogation and bigoted perspectives, I think we should hope for a tighter connection here between illocution and cognition.

This leads me to a different concern that Hom does not press. If we locate the derogatory content of slurs in the illocutions they help to constitute, then we suppose that the slur is only problematic when it is incarnated in a speech act. But consider a bigot who does not slur, but nonetheless slurs feature regularly in their internal monologue and beliefs–a person who consistently conceptualizes targets of slurs under the aspect of those slurs. There is something derogatory–something nasty–in his thoughts, but that something may never achieve illocutionary fruition. We should want to say that, even if to a lesser degree, the offensive content of the slur still appears in his head. And when that bigot eventually does slur, some of the offensive content of his slur is continuous with the offensive content of the thoughts which contained it. It is not as though he has thought a neutral thought and it has come out as an act of derogation. His words have expressed a piece of him; his thoughts, his cares, his world. We can often forgive poor behavior, but the casual use of a slur can cause the reevaluation of a friendship; a shunning or an exile. Hornsby’s gestural view says little about this revelatory feature of slurs. We may not wish to locate the derogatory content of a slur neatly in its semantic, truth-conditional content, but nonetheless gestures do not seem cognitive enough.

Hom also argues that these posited gestures are mysterious. One does not actually gesture when they use a slur, so what is this imaginary, immaterial gesture? It is clear to me that Hornsby refers to gestures as a helpful metaphor (e.g., “it is as if” and “is to be thought of as if it were…”), and not quite as a theory. I expect—especially given that this metaphor takes up one single paragraph of the paper—Hornsby would welcome a filling in of this metaphor into a more determinate theory. As I have said above, Liu’s illocution indication view is a plausible continuation of Hornsby’s research programme.

The last objection from Hom I’d like to consider is that Hornsby fails to accommodate an important feature of several slurs: appropriated usages. For example, if the n-word were akin to a neutral counterpart with an accompanying derogatory gesture, then how does it find such casual, even positive usage within black communities? Hornsby does consider words whose derogatory force has been reduced, if not eliminated, over time, but she does not consider words which have both ubiquitous derogatory and appropriated usages simultaneously (e.g., the n-word, ‘queer’, ‘bitch’). Here is one option for Hornsby: derogation, like any other illocutionary act, has felicity conditions. It does not count as apologizing to John if I utter the right words while alone. Perhaps, then, it does not count as derogating John if the one who utters the slur does so in the wrong sort of context. This strategy seems promising, but Hornsby owes us an account of the felicity conditions of slurs. This is not an easy task, as her felicity conditions must preclude appropriated usages from derogating while maintaining that usages by the theorist do derogate.

These objections have yielded some tasks for Hornsby and her followers. A fleshed out version of her theory must (i) capture the impression that the user of a slur has gotten something wrong, (ii) demystify the metaphor of gesture, and (iii) accommodate appropriated uses while excluding theoretical ones.

Conclusion

Despite its obscurity and heterodoxy, or perhaps exactly because of these features, Hornsby’s work on the meanings of slurs is incredibly far-reaching and provocative. Nonetheless, as Hom allows us to see, it leaves much to be desired. It remains to be seen whether an account can be developed that is both faithful to Hornsby’s intentions and satisfies the above desiderata. I am optimistic, and the purpose of part 2 of this post will be to explore such a possibility.

References

Camp, Elisabeth (2018). “A Dual Act Analysis of Slurs”. In David Sosa, Bad Words: Philosophical Perspectives on Slurs. Oxford, United Kingdom: Oxford University Press. pp. 29-59.

Cavell, Stanley (1976). “Knowing and Acknowledging”. In Cavell, Stanley Must we Mean What we Say? Cambridge University Press. pp. 238-266.

Hornsby, Jennifer (2001). “Meaning and uselessness: how to think about derogatory words”. Midwest Studies in Philosophy 25 (1), pp. 128-141.

Lasersohn, Peter (2007). “Expressives, Perspective, and Presupposition,” Theoretical Linguistics 33: 223–30.

Liu, Chang (2020). Slurs as Illocutionary Force Indicators. Philosophia 49 (3):1051-1065.

McDowell, John. (2001). “In Defense of Modesty”. In McDowell, John Meaning, Knowledge and Reality. Harvard University Press.