On Slurs

Introduction

In this essay, I argue that racial epithets cannot be understood through combinatorial externalism. Furthermore, if racial epithets derogate through a speech-act, this shows that racial epithets derogatory content must be pragmatic. Therefore, racial epithets should be understood through a pragmatic strategy.

The body of this essay consists of two sections. In Section one, I explain the fundamentals and evaluate strategies for understanding slurs. I also create criteria for what an adequate strategy for understanding racial epithets should include. In Section two, I examine 'combinatorial externalism'. I argue that this strategy fails, and that racial epithets cannot be understood semantically.

Finally, I conclude that Hom is mistaken in asserting combinatorial externalism is the correct way of understanding racial epithets. The derogatory content of racial epithets, thus racial epithets themselves, can only be understood pragmatically. I also explain what conditions would prove my claim wrong.

I. Semantic and Pragmatic Strategies

'Racial epithets' (slurs) are derogatory insults which apply to members of a race. Hom (2008) outlines the two branches of theories which claim to explain how slurs should be understood. These are semantic strategies, and pragmatic strategies. Hom asserts that, despite pragmatic strategies seeming more intuitive, they do not help to understand slurs (Hom, 2008, p. 416).

Semantic strategies assert that the derogatory content of slurs is fundamental to their meanings (Hom, 2008, p. 416). Therefore, this content is always expressed when one utters a slur (Hom, 2008, p. 416). Furthermore, since it is always expressing derogatory content, context cannot affect derogatory meaning (Hom, 2008, p. 416).

On the other hand, pragmatic strategies claim that the context of a slur's utterance changes its meaning. Therefore, slurs do not have a fixed meaning, nor are they always derogatory. It matters how they are used, and who they are used by and on. In summary, context must be accounted for to understand racial epithets. (Hom, 2008, p. 416)

Hom asserts that a successful semantic strategy must describe the semantic content of racial epithets while allowing for both the shifty and scalar aspects of epithets (Hom, 2008, p. 418). A slur is 'shifty' if it only sometimes derogates, like in appropriated usage (Hom 2008, pp. 418-419). Scalar refers to the difference in power between slurs. Hom gives the example of the 'n-word' having more derogatory force than 'chink' (Hom, 2008, p. 419).

Furthermore, an explanation must not posit "mysterious, ad hoc entities," (Hom, 2008, p.418). Silentism fails to do this (Hom, 2008, p. 417-419). Silentism is a semantic strategy which asserts that uttering an epithet is equivalent to stating a 'non-pejorative correlate' for the same race (i.e., black, African American), while simultaneously performing a negative gesture. This could be displaying middle-fingers, or performing a throat slashing gesture (Hom, 2008, pp. 417-418). The gestures are mysterious since their meaning is unclear, especially since it is unclear which should be used. Middle fingers are merely rude, but throat slashing is threatening. Furthermore, Hom asserts they are ad hoc since they are created in the first place to explain epithets (Hom, 2008, p. 418).

Additionally, silentism cannot account for the scalar aspect of slurs. Since slurs are scalar (Hom, 2008, p. 419) it is unclear how a non-pejorative correlate with a gesture could explain the differences in derogatory force between the 'n-word' and 'chink'. If this strategy accepted pragmatic elements, this could be remedied by asserting that racist language or behaviour directed towards a black person is a worse act than towards a Chinese person, because of historical oppression. This is dependent on the racial history of where this is uttered, and of who utters it.

Hom also argues pragmatic strategies fail to provide understanding of racial epithets (Hom, 2008, pp. 419-426). The three strategies he critiques - Radical contextualism, Fregean minimalism, and Pragmatic minimalism - each fail in different ways. Radical contextualism fails to provide a framework for deriving the content of an epithet from the contextual features where it is uttered (Hom, 2008, p. 419). This means that there is no way to predict what a slur will mean, since the meaning is only clear in the specific context it is uttered in.

Fregean minimalism also lacks an explanation of what context features matter for a slur's meaning. However, it doesn't make the same assertion that all meaning is contextual. Instead, it derives understanding from Frege's (1948) ideas of sense and colouring. 'African American' and the 'n-word' share the same sense in that they refer to the same group - black people - but the shading of the words creates a different tone (Frege, 1948, p. 213). 'African American' has a neutral tone, it is not praising nor derogatory, whereas the 'n-word' has a negative tone, since it is derogatory (Hom, 2008, p. 420). Hom asserts that Frege's cannot be applied here so freely, since it lacks explanation of what context pieces will colour these epithets (Hom, 2008, p. 421). Therefore, an additional framework must be provided. Furthermore, the subjective nature of colouring does not align with the near-universal understanding of the force of the n-word (Frege, 1948, p. 213; Hom, 2008, pp. 420-421). In other words, we would expect more variation in interpretation of the word if Fregean minimalism were correct.

Finally, Hom evaluates pragmatic minimalism. This view states that derogatory content is specifically determined in each context through conventional implicature (Hom, 2008, p. 423). This means that derogation is implied, rather than being a part of what is strictly and literally said (Grice, 1975; Hom, 2008, p. 423). Hom asserts that this fails to capture the internal meaning of slurs. Therefore, slurs are identical with their non-pejorative counterparts (literally, the 'n-word' is equivalent to 'black'). Furthermore, if the derogatory aspect is conventionally implied, then derogation should occur in every context (Hom, 2008, p. 424). However, non-derogating uses of slurs exist (Hom, 2008, p. 424). Therefore, this view fails to give an adequate understanding of slurs.

These criticisms can be summarised to provide criteria for what a strategy for understanding slurs should be. A working strategy is:

1. Not mysterious.
2. Not ad hoc.
3. Capable of explaining how slurs are shifty and,
4. Scalar.
5. Capable of pointing out which context features affect meaning and derogation, and
6. How they effect meaning and derogation.

II. Combinatorial Externalism

Hom argues his strategy, combinatorial externalism (CE), is the strategy that should be used to understand racial epithets. (Hom, 2008, p. 416). Therefore, CE should, at minimum, meet the criteria outlined.

CE claims that the derogatory meanings of epithets are determined semantically by social institutions of racism (Hom, 2008, p. 430). These institutions have two parts. The first is ideology: negative beliefs about a race (Hom, 2008, p. 431). The second is practice: the mistreatment of a race (Hom, 2008, p. 431). When put together, the two reinforce one another. Thus, slur meaning can be described as essentially saying that:

The members of x race ought to be discriminated against. This is because they possess supposed negative properties, by virtue of being members of x race (Hom, 2008, p. 431).

Hom asserts that this meaning is semantic, despite being external (Hom, 2008, p. 430; Kripke, 1980). In other words, its meaning is contained within the word, despite being brought from an external entity.

Finally, Hom asserts that epithets always express derogatory semantic content, but do not always derogate their targets (Hom, 2008, p. 432). Hom states the derogation itself is a type of speech-act (Hom, 2008, p. 432; Austin, 1962). Therefore, to derogate, one must apply the epithet to a target (Hom, 2008, p. 432). In doing this, the speaker asserts their target has negative properties because they are of that race, thus they deserve mistreatment (Hom, 2008, p. 432). Essentially, epithets have derogatory content, and speakers derogate when uttering epithets at a target (Hom, 2008, p. 432).

This strategy avoids some problems of other strategies. First, it explains how the semantic content should be understood. Firstly, unlike silentism, it does not invent an entity to explain racial epithets, since it invokes social institutions of racism (Hom, 2008, p. 431). Additionally, social institutions of racism are not mysterious: anyone can point to the ideology and practice of racism in some form, such as the false beliefs that different races have different IQs (Templeton, 2001), or racially motivated killings, such as the murder of Emmett Till (Library of Congress, 2019).

Next, CE also seems to explain how slurs can be shifty and scalar. For shifty, the fact that it derogates sometimes and not others is explained by derogation only occurring as a speech-act (Hom, 2008, p. 432; Austin, 1962). Therefore, someone can use it without performing the speech-act of derogation (Austin, 1962). For scalar, the derogatory force of each epithet is be determined externally by the social institutions of racism (Hom, 2008, p. 431). Therefore, the semantic, derogatory force of an epithet can be found in how hateful the ideology is, and how adversely the practice of racism effects the race corresponding to an epithet.

However, CE fails to adequately explain which context features affect meaning and derogation, beyond simply 'social institutions of racism'. This affects how well it accounts for the shiftiness and scaling nature of slurs.

One instance where it is unable to explain the impact of these features is in nonderogatory, nonappropriated (NDNA) contexts. Since Hom labels derogation as a speech-act done towards a subject, he is forced into the conclusion that phrases like "Yao Ming is Chinese, but he's not a chink" are not derogatory, since it is not derogating the subject, Yao Ming, with the word 'chink' (Hom, 2008, p. 438).

However, this is hard to defend. If the word is defined by a social institution of racism, uttering this word should still imply that Chinese people ought to be discriminated against, because their race is deserving of it (Hom, 2008, p. 431). The utterance of "Yao Ming is Chinese, but he's not a chink" (Hom, 2008, p. 438), seems only to insist that some Chinese people are not a part of the Chinese, plus 'chink' category. In other words, while it does not, as a speech-act, derogate Yao Ming specifically, it still derogates the rest of the Chinese people, who are not excluded from being 'chinks'.

Hom partially answers this criticism in his later paper 'Pejoratives as Fiction' (Sosa, 2018). Simplifying massively, he asserts that racial epithets function like mythology. In other words, 'chinks' are real only in the same way 'unicorn horns' are real. That is, that they have an accessible (though, technically incorrect) referent, and that while it is materially false, it is simultaneously mythologically true that Chinese people are 'chinks' (Sosa, 2018, p. 110), at least, according to the mythos created by a social institution of racism. Therefore, this is not an NDNA use, since uttering it asserts a racist mythos upon Chinese people.

Speech-acts also cause issues for CE. By invoking speech-acts, the conclusion should be that derogatory variation is determined pragmatically, not semantically. Hom claims that derogatory variation is determined by a social institution of racism, so derogatory variation appears because of different attitudes to different races. In other words, in what is strictly and literally said (Grice, 1975). This leaves no room for implication, nor for speech-acts, since, as mythologisation has shown, the mere uttering of these words would seem to imply derogation. And yet, there must be room for it, since appropriation requires a non-derogatory use.

Furthermore, there are more categories than oppressor and oppressed that must be explained. There is not one uniform, nor averageable level of racism per race, as Hom implies. Contrarily, if one holds that racism is prejudice plus power (Bidol, 1972), then in some contexts, racial epithets seem to lose derogatory force entirely. While Hom mentions this in the context of words losing their semantic, racist, meanings over time, I argue that these two meanings exist simultaneously.

A Chinese person called a 'chink' by a white man in the UK faces both power and prejudice in the systemic racism of white supremacy. But a Chinese person being called a 'chink' by a non-white, poor immigrant in China fails to have much derogatory force, since there is no inherent devaluation in being a 'chink' in China: they do not have the same social institution of racism. The speech-act therefore fails, and so doesn't derogate.

This does not endorse Hom's conclusion that derogation isn't semantic but derogatory content is. Without the semantically defining social institution of racism, derogation can't occur because it lacks derogatory content in this context. Since it can be contextually lacked, it cannot be semantic.

Conclusion

Racial epithets cannot be understood semantically. While the pragmatic and semantic strategies discussed are genuinely problematic, CE also fails according to some of the same criteria, and additional considerations. Speech-act theory combined with semantic externalism shows that racial epithets must be understood pragmatically. This is evidenced by the speech-act of derogation failing specifically because there is no derogatory content to be found in the word. This is the result of semantic externalism and the social institution of racism. If the exact same word in the exact same language has different content at the exact same time in two different places, or with two different speakers, then this meaning must be considered pragmatic, since it is clearly not determined intrinsically.

To prove my argument wrong, it must be shown that the racial epithet spoken is a genuinely different word despite its similarities. This would allow it to have a different semantic meaning. This seems to require proving that there are many social institutions of racism, each determining the semantic meaning of words independently. This would look like two different social institutions of racism independently producing two words, both literally being the 'n-word' in every visible aspect, but with different semantic meanings. To me, this seems unlikely, or, at very least, hard to comprehend.

Bibliography

Austin, J.L. (1962) How to Do Things with Words. Oxford University Press.

Bidol, P. A. (1972). 'Developing New Perspectives on Race: An Innovative Multi-media Social Studies Curriculum in Racism Awareness for the Secondary Level.' New Perspectives on Race.

Frege, G. (1948) 'Sense and Reference', The Philosophical Review, vol. 57, no. 3, 1948, pp. 209–30. JSTOR, https://doi.org/10.2307/2181485. Accessed 02-05-2026.

Grice, H.P. (1975) ‘Logic and conversation’, Communications, 30 (1), pp. 41–58. Available at: https://doi.org/10.3406/comm.1979.1446.

Hom, C. (2008) 'The Semantics of Racial Epithets', The Journal of Philosophy, Vol. 105, No. 8, pp. 416-440. Available at:
https://www.jstor.org/stable/20620116 (Accessed: 02-05-2026).

Hom, C and May, R. (2018) 'Pejoratives as Fiction', in Sosa, D. (ed.), Bad Words: Philosophical Perspectives on Slurs, Engaging Philosophy (Oxford, 2018; online edn, Oxford Academic, 23 Aug. 2018), pp. 108-131. Available at:
https://doi.org/10.1093/oso/9780198758655.003.0006 (Accessed: 02-05-2026)

Kripke, S.A. (1980) Naming and Necessity. Harvard University Press.

Library of Congress (2019). The Murder of Emmett Till  | Articles and Essays  | Civil Rights History Project  | Digital Collections  | Library of Congress. [online] The Library of Congress. Available at: https://www.loc.gov/collections/civil-rights-history-project/articles-and-essays/murder-of-emmett-till/. (Accessed: 03-05-2026)

Templeton, R. (2001), 'The Genetic and Evolutionary Significance of Human Races', in Fish, J. M. (ed.), Race and Intelligence Separating Science from Myth, London: Routledge, pp. 31–55.