Professional Documents
Culture Documents
DOI 10.1007/s11098-014-0338-4
Abstract What accounts for the offensive character of pejoratives and slurs,
words like ‘kike’ and ‘nigger’? Is it due to a semantic feature of the words or to a
pragmatic feature of their use? Is it due to a violation of a group’s desires to not be
called by certain terms? Is it due to a violation of etiquette? According to one kind
of view, pejoratives and the non-pejorative terms with which they are related—the
‘neutral counterpart’ terms—have different meanings or senses, and this explains
the offensiveness of the pejoratives. We call theories of this kind, semantic theories
of the pejoratives. Our goal is broadly speaking two-fold. First, we will undermine
the arguments that are supposed to establish the distinction in meaning between
words like ‘African American’ and ‘nigger’. We will show that the arguments are
suspect and generalize in untoward ways. Second, we will provide a series of
arguments against semantic theories. For simplicity, we focus on a semantic theory
that has been proposed by Hom (J Philos 105:416–440, 2008) and Hom and May
(Anal Philos 54:293–313, 2013). By showing the systematic ways in which their
view fails we hope to provide general lessons about why we should avoid semantic
theories of the pejoratives.
1 Introduction
And to see the cool way of that nigger—why, he wouldn’t a give me the road
if I hadn’t shoved him out o’ the way. I says to the people, why ain’t this
nigger put up at auction and sold? that’s what I want to know. (Mark Twain,
Huckleberry Finn)
123
A. Sennet, D. Copp
There’s one term in the above quotation that probably stands out as particularly
offensive, that you probably would not say out loud in public, that you may well be
embarrassed to have people see you read. Words like ‘nigger’ are known as
‘pejoratives’ (or ‘slurs’ or ‘derogatory terms’) on account of their offensive
character.1 While it is not clear that they form anything like a natural linguistic
class, they raise an interesting question—what accounts for their familiar offensive
character? Is it due to a semantic feature of the words, or, perhaps, a pragmatic
feature of their use? Is it due to a violation of a group’s desires to not be called by
certain terms? Is it due to a general convention that governs the terms’ use? Is it a
matter of etiquette that certain words are not to be used?
Here’s one line of thought that some philosophers have found highly persuasive.
The difference between words like ‘nigger’ and ‘black’, if not semantic in character,
should make no difference to the truth conditions of the sentences in which they
appear. Substitutions of one for the other, however, can make a semantic difference
to the sentences in which they appear—it appears that they can alter the truth value
of a sentence in which they appear. Therefore, the difference between the words is
semantic in character. That is, generalizing, pejoratives and the non-pejorative terms
with which they are related—which we will refer to as their ‘neutral counterparts’—
make different contributions to the content of the propositions expressed by
sentences in which they appear.
The argument, even if sound, is compatible with the view that words like ‘nigger’
do not contribute semantically to the content of propositions expressed by sentences
in which they appear. Some have argued that pejorative terms have an expressive
function but make no semantic contribution. For the purposes of this paper,
however, we will assume that pejoratives do make a semantic contribution to the
sentences in which they appear. Armed with this assumption, those who accept the
conclusion that pejoratives differ semantically from their neutral counterparts can
focus on how they differ semantically. How are ‘nigger’ and ‘black’ different? Well,
it is pretty clear that no-one can satisfy the term ‘nigger’ without being black. But if
‘nigger’ differs semantically from ‘black’, there must be more to the meaning of the
former. And presumably the extra ingredients of meaning should help explain why
terms like ‘nigger’ are offensive. This defines a research program into the semantics
of pejoratives. Given a pair of terms, one of which is pejorative and one of which is
its neutral counterpart, what is the extra ingredient in the meaning of the pejorative
that helps to explain its offensiveness? We will call theories that aim to explain the
offensiveness of pejoratives in this way, by proposing an extra ingredient in the
meanings of pejoratives, ‘semantic theories’ of the pejoratives.
We will discuss a variety of semantic theories since our interest is not with any
particular theory but with the general class of semantic theories. Our goal is broadly
1
We use these three terms interchangeably in accord with definitions given in the American Heritage
Dictionary (Third Edition). It defines a ‘pejorative’ as a ‘disparaging or belittling word or expression’ and
‘slur’ in the relevant sense as ‘a disparaging remark’. It defines ‘derogatory’ as ‘disparaging or belittling’,
so that a derogatory term would be a disparaging or belittling term. An anonymous referee pointed out
that some philosophers seem to draw a distinction between pejoratives and derogatory terms. We have no
objection to using the terms to draw finer distinctions, but we treat them as equivalent.
123
What kind of a mistake is it to use a slur?
speaking two-fold. First, we will undermine the arguments that are supposed to
establish the distinction in meaning between words like ‘black’ and ‘nigger’. We will
show that the relevant arguments are suspect and generalize in untoward ways.
Second, we will go on the offensive and provide arguments against semantic views of
pejoratives. For simplicity, we will focus on a semantic theory of pejoratives that has
been proposed by Hom (2008) and Hom and May (2013).2 We hope that showing the
systematic ways in which their view fails provides general lessons about why we
should avoid semantic theories. In this way, we hope to undermine the appeal of such
theories. We thus hope to make the field safe again for non-semantic views of
pejoratives (which we both think are plausible, though we differ on which one is true).
To clarify, we use the term ‘semantics’ in a narrow way such that the semantics of a
term determines its contribution to the truth-conditional content of sentences in
which it appears.3 A semantic theory of the pejoratives seeks to explain the
offensiveness of pejoratives on the basis of truth-condition affecting properties of
pejorative terms, properties the terms have due to linguistic conventions governing
their use. Non-semantic theories seek to explain the offensiveness of pejoratives in
some other way, such as by reference to pragmatic phenomena or by reference to
usage conventions that do not affect truth-conditional content. Accordingly,
semantic theories of the pejoratives have two central features. (a) They posit a
distinct truth-condition-affecting-meaning for pejorative terms that distinguishes the
semantics of these terms from the semantics of their neutral counterparts. (b) They
use that distinct meaning to explain the offensive character of pejorative terms.4
It is useful to distinguish between semantic theories that claim or imply that, for
instance, ‘black’ and ‘nigger’ differ in extension and theories that do not claim or
imply this. For example, a theory could assign to ‘nigger’ the same meaning as it
assigns to ‘black for whom white racists have contempt’.5 This would perhaps
explain the offensiveness of the term and it would seem to be true on this account
that ‘black’ and ‘nigger’ have the same extension. Some semantic theories imply,
2
See Richard (2008) for a semantic view that treats (many) sentences with pejorative terms as lacking
truth values altogether.
3
Potts (2005) distinguishes between ‘at-issue’ truth-conditional content and ‘not at-issue’ content. In
these terms, we construe semantics in the narrow sense to be concerned with ‘at-issue content’.
4
Miščević (2011), Richard (2008), Croom (2013) and Bach (2014) defend semantic accounts. Jesion
(2013) presents an account that is not semantic by our lights and Camp (2013) offers a perspectival
account that we believe doesn’t count as semantic by our lights. Anderson and Lepore (2013a) offer a
non-semantic theory of the pejoratives in which the offensive character of pejoratives is explained by a
general prohibition against using them. Thanks to an anonymous referee for helping us clarify this point.
Hom (2010) offers a helpful and comprehensive taxonomy of views regarding pejoratives.
5
This is just a toy example of a theory of the meaning of ‘nigger’. An anonymous referee pointed out
that it is not necessarily the case that ‘nigger’ and ‘black’ have the same extension on this theory since
some racists might lack contempt for some black people. There are more subtle ‘same extension theories’
one could imagine but we do not pursue the issue here.
123
A. Sennet, D. Copp
however, given plausible assumptions, that ‘nigger’ and ‘black’ differ in extension.
Consider for example a theory that assigns to ‘nigger’ the same meaning as it
assigns to ‘Black person who deserves contempt for being black.’6 This account also
would explain the offensiveness of the term, but given that no-one deserves
contempt for being black, it implies that ‘There are no niggers’ is true. We will
discuss semantic theories of both kinds, which we will call ‘same extension’ or SE
theories and ‘different extension’ or DE theories.
Semantic theories use features of a standard truth-conditional semantics in
providing an account of the offensive character of pejoratives. For example, a theory
that builds on a Fregean semantics would treat the offensive character of pejoratives
as due to either their sense or their sense and reference. A non-Fregean theory might
treat the offensive character of pejoratives as due to the nature of the presumably
complex properties they ascribe. In one way or another, semantic theories propose
that the offensive character of the terms is due to an aspect of their standard semantics.
Pejorative terms are typically associated with neutral counterpart terms in the
way that ‘kike’ is associated with ‘Jew’ and ‘nigger’ with ‘black.’ Semantic views
can be characterized as treating the semantics of a pejorative as some function of the
semantics of its neutral counterpart. So if ‘Jew’ ascribes the property of being
Jewish, ‘kike’ might be thought to ascribe a related corresponding complex
property, such as, perhaps, the property of being Jewish and contemptible in virtue
of being Jewish. In general (following Hom and May 2013) we might use the label
‘PEJ’ to refer to the function that, according to a theory, takes the semantic value of
a non-pejorative to the semantic value of the corresponding pejorative. On one
proposal, for example, PEJ takes the property that is the semantic value of a non-
pejorative to the conjunction of that property and the property of being contemptible
in some way related to having the first property.7 Semantic theories differ in the PEJ
function that they propose.8 Given what we have said, semantic views imply that the
truth conditions of typical sentences containing a pejorative differ from the truth
conditions of the corresponding sentences that differ only in that the neutral
counterpart replaces the pejorative. So, for instance, on the proposal sketched above
‘Smith is a kike’ differs in truth conditions from ‘Smith is a Jew’ since, unlike the
latter sentence, it entails that Smith is contemptible in virtue of being Jewish.
Just as there is a variety of semantic theories, there is a variety of non-semantic
theories of pejoratives. For example, on the view proposed by one of us, there are
usage conventions governing the use of pejoratives such that, roughly, if ‘P’ is a
pejorative, it is linguistically inappropriate to call someone ‘a P’ unless one has
6
For a similar view, see Bach (2014).
7
It’s worth noting that some pejorative terms are narrower in presumed extension than the presumed
non-pejorative counterpart. Chris Rock’s ‘Niggas vs. Black People’ gives an example of trying to narrow
the extension of ‘nigga’ by comparison with the extension of ‘African American’. On his account, the
non-pejorative counterpart of ‘nigga’ might be a highly complex property (involving being African
American, shooting at movie screens, and so on, for example).
8
Moreover, as Hom (2008) points out, PEJ will have to be fairly complex in nature if it is to be able to
capture the differences in pejorative terms that are generated from the same non-pejorative counterpart
(i.e. ‘Heeb’ seems to be less pejorative in nature than ‘Kike’). Non-semantic views will have to explain
degrees of pejoration as well.
123
What kind of a mistake is it to use a slur?
contempt for that person or for members of the relevant group (Copp 2001, 2009).
One would not be competent with such a term if one did not know this. Similarly, to
use an example suggested by Kent Bach, there is a usage convention governing the
use of the second person pronoun ‘tu’ in French to the effect that one uses it, rather
than ‘vous’, only to address people with whom one is on familiar terms. In virtue of
this convention, when someone addresses another person as ‘tu’, she conventionally
implicates that she is on familiar terms with that person. Similarly, in virtue of the
nature of the usage conventions governing the use of pejoratives, a person who calls
someone ‘a P’, using a pejorative term ‘P’, conventionally implicates that she has
contempt for members of the relevant group. This is only one example of a non-
semantic view.
Hom (2008) Hom and May (2013) offer several arguments against the
conventional implicature view, a view that Hom calls ‘pragmatic minimalism’
(2008). One of us has previously responded to some of these arguments (Copp
2009) and we offer novel responses to some of their arguments below. Our goal,
however, is to answer Hom and May’s more general arguments against non-
semantic views and in favor of semantic views.
The main difference between semantic and non-semantic views of the
pejoratives, then, is that semantic views, unlike non-semantic views, imply that
the truth conditions of typical sentences containing pejoratives differ from the truth
conditions of corresponding sentences in which the neutral counterpart replaces the
pejorative. This difference is the focus of some of the arguments we will discuss.
Hom and May propose a semantic DE view that is noteworthy because, as they
interpret it, it entails that ‘There are no kikes’ and ‘There are no niggers’ are
necessarily true. In a canonical statement of their view, they say,
The semantic interpretation of PEJ(x) is neatly expressible in Fregean terms.
The sense expressed by PEJ denotes a second-level function that combines
with a first-level concept (e.g. of race, gender, religion or class) to form a
complex first level pejorative concept. The complex pejorative concept takes
objects as inputs, and has falsity as its output. That complex pejorative
concepts are constant functions mapping individuals to falsity is the
realization of the null extension thesis. Intensionally, we can think of PEJ
as a concept abstracted from the moral truth that nobody ought to be the target
of negative moral evaluation because of being Jewish; that is, the concept x
ought to be the target of negative moral evaluation because of being x. This
concept can be abstracted from any truth of this form,…; accordingly, each of
the distinct instances of PEJ(x) will correspond to a distinct pejorative and
each will have a null extension. (Hom and May 2013, pp. 298–299)
Basically, the account explains the semantic value of a pejorative as the result of
applying the function PEJ to the semantic value of its neutral counterpart. So if the
semantic value of ‘Jew’ is the concept of being Jewish, the semantic value of ‘kike’
123
A. Sennet, D. Copp
123
What kind of a mistake is it to use a slur?
are, any existing institutions that discriminate against Caucasians. In any case, in
what follows we set aside Hom’s original account and instead discuss Hom and
May’s (2013) account. It will serve as our example of a semantic theory, and, more
specifically, of a DE theory.
One might quibble with their account. Murderers are those who kill people
without sufficient moral justification, and they deserve negative moral evaluation on
this account. Moreover, we can stipulate that the concept of murderer* is the concept
derived from ‘kill without sufficient moral justification’ by the application of PEJ. It
is therefore the concept of being an appropriate target of negative moral evaluation
on account of having killed without sufficient moral justification. On Hom and May’s
account, then, it would seem that ‘murderer*’ may count as a pejorative. If so, this is
a counter-example to their view. It would be a counter-example on two counts. First
‘murderer*’ has an extension. There are people who deserve negative moral
evaluation on account of having killed without sufficient moral justification. Second,
‘murderer*’ does not actually seem to be a pejorative. It is not a disparaging or
belittling term and it is not offensive to call someone a ‘murderer*’ if the person has
indeed killed someone without sufficient moral justification. So if ‘murderer*’ is a
pejorative, Hom and May’s null extension thesis is false. If ‘murderer*’ is not a
pejorative, then their account of the semantics of pejoratives is too broad. It
incorrectly implies that some non-pejoratives are pejorative.
Hom and May seem to stipulate, however, that the non-pejorative terms that
correspond to pejoratives express concepts ‘‘of race, gender, religion or class.’’ They
might deny on this basis that the concept of murderer* can be derived from ‘kill
without sufficient moral justification’ by application of PEJ. It is hard to see the
motivation for this stipulation, however, and there seem still to be counterexamples.
For consider the concept of being a member of the exploiting ruling class. Since this
is a concept of class, it would seem that PEJ would take it to a concept, which we
stipulate to be expressed by the term ‘exploiter*’, viz., the concept of being an
appropriate target of negative moral evaluation on account of being a member of the
exploiting ruling class. Yet the term ‘exploiter*’ intuitively is not a pejorative even
though it should be one on Hom and May’s account. It is not a disparaging or
belittling term and it would not be offensive to call someone who is a member of the
exploiting ruling class an ‘exploiter*’. Moreover, it is true, or at least arguable, that
there are exploiters*, for it is arguable that people who are members of the exploiting
ruling class are appropriate targets of negative moral evaluation on this account. So if
‘exploiter*’ is a pejorative, Hom and May’s null extension thesis is false. If
‘exploiter*’ is not a pejorative, then their account of the semantics of pejoratives is
too broad. It incorrectly implies that some non-pejoratives are pejorative.
Hom and May might modify their account in response to this objection. They
might stipulate, for example, that a term (PEJ)N qualifies as a pejorative only if it is
not the case that anyone would be an appropriate target of negative moral evaluation
on account of being N.9 So understood, the null extension thesis for pejoratives
follows from their account. Moreover, on this understanding, terms like ‘murderer*’
9
See Hom and May 2013, p. 299, note 15, where they seem to make this stipulation.
123
A. Sennet, D. Copp
and ‘exploiter*’ are not pejoratives. Even given this stipulation, however, their
account is liable to counter-examples of a different kind. For there are pejorative
terms that don’t seem to involve any suggestion of moral evaluation. ‘Retard’,
‘dwarf’ and certain other pejorative terms don’t bring with them any sense of moral
evaluation, though they are offensive, disparaging and belittling terms to use.10 We
return to issues of this kind later in the paper.
Hom and May’s semantic DE theory illustrates a constraint that we think any
plausible semantic theory must meet. An anti-Semite uses ‘kike’ to pejorate Jews,
and in general, a bigot uses a pejorative PEJ(N) to pejorate those who are N.
Therefore, plausibly, either PEJ(N) and N have the same extension or a bigot who
uses PEJ(N) would believe that PEJ(N) and N have the same extension. Moreover,
since bigotry does not generally introduce semantic ignorance, we think a plausible
DE theory should not imply that a person who believes that PEJ(N) and N have the
same extension would thereby be making any kind of semantic error. Hom and
May’s theory meets this constraint, since on their view, the error a bigot makes in
thinking that, e.g., ‘kike’ and ‘Jew’ have the same extension, is a moral error, not a
semantic error.
From the perspective of racists and bigots, Hom and May’s account will seem to
be too good to be true. Racists and bigots might hope that the world is such that
‘There are no kikes and no niggers’ is true, and they might have wished that there
are no possible worlds where there are Jews and blacks. But they are not likely to be
relieved to be told by Hom and May that in fact there never was any possibility of
there being people who could qualify as ‘kikes’ or ‘niggers’. We need to look into
the reasons Hom and May give for their account and into the objections it faces. We
also need to bear in mind, of course, that their view is simply one example. Other
semantic DE theories, or some SE theory, might be able to avoid some of the
problems faced by Hom and May.
10
Thanks to Kent Bach for discussion and examples.
123
What kind of a mistake is it to use a slur?
Intuitively, the corresponding odd and even numbered sentences in the list have
different meanings. Intuitively, (1) is perhaps true, but even if (1) is true, (2) is false.
(3) is false but (4) is true. Hom and May conclude that we ought to reject any view
that takes the expressions ‘black’ and ‘nigger’ to be synonymous. Mutatis mutandis
for other pejorative expressions such as ‘kike’, ‘chink’, and ‘wop’. Hom and May
conclude that the pejorative terms do not mean the same as the corresponding
neutral counterparts.
4.2 Conceivability
123
A. Sennet, D. Copp
If this argument is invalid—and Hom and May are committed to its invalidity—then
presumably ‘nigger’ is contributing content to (12) that is not present in the
conjunction of (10) and (11). But this means that ‘nigger’ contributes semantic
content to the truth conditions of (12) and that this content differs from the semantic
content contributed by ‘black’. Therefore, ‘nigger’ doesn’t have the same meaning
as ‘black’.
11
In conversation, Kent Bach suggested to us that perhaps the notion of convention typically at play is a
misleading one to apply to cases such as ‘but’. We aren’t really invested in the name or any particular
theory of how the contrastive nature of ‘but’ claims arise so long as they are, as per orthodoxy, not truth-
conditional effects.
123
What kind of a mistake is it to use a slur?
12
Hom agrees with Bach (1999) that there are no conventional implicatures. We disagree. One of us has
previously discussed some of Bach’s arguments (Copp 2009).
123
A. Sennet, D. Copp
Hom and May’s conceivability argument starts with a premise about what is
conceivable:
(31) It’s conceivable for there to be Jews without kikes.
Hom and May’s defense of (31) is that in a possible world where there is no racism
but where there are Jewish people, it would be conceivable for there to be Jews but
no kikes. We find this reasoning very puzzling. Why should the existence of racism
have any bearing whatsoever on the truth of (31)? We suspect that Hom and May
might have confused what we mean by our terms when we consider other possible
worlds with what people at those worlds would mean using those same terms. That
is, they might have confused the determination of the meaning of the word ‘kike’ at
the actual world and its application to an anti-Semitism-free world with whether or
not the word ‘kike’ would have meant the same thing at an anti-Semitism-free
world. Another possibility is that they might have confused whether a world is
conceivable with whether the people in a conceivable world would claim that
123
What kind of a mistake is it to use a slur?
Sometimes fights over inferences come down to whether one is willing to assert a
conclusion, given that one accepts some premises, rather than whether one thinks
the conclusion follows validly from the premises. But these are two different
matters. The following example illustrates the point. We are perfectly willing to
accept (32) but unwilling to assert the conclusion (33):
(32) Obama is black and smart.
(33) Obama is black but smart.
We are unwilling to assert (33) despite believing that ‘but’ and ‘and’ are
synonymous and differ only in conventions of use. This is because, on the standard
account, a felicity condition for asserting (33) is believing that there is a contrast
between being black and being smart, which we don’t believe. We thus can’t
felicitously assert (33) even though we think the proposition it expresses is true.
It is a delicate matter whether we reject an inference because it may lead from a
true premise to a false conclusion or whether instead we reject the inference because
it proceeds from a true premise to a conclusion that is infelicitous to assert.
Nonetheless, it is clear that inferences may fail to preserve felicity even if they
preserve truth. This gives us an explanation for the failure of the inference that
bothered Kaplan. One won’t draw the inference because the addition of some words
leaves the proposition expressed unchanged but not the felicity conditions for
asserting the sentence that expresses that proposition. Thus, I certainly may accept
(34) and (35) but refuse to accept the conclusion that (36):
(34) Obama won the election.
(35) Obama was black.
(36) That nigger Obama won the election.
The refusal to assent to sentence (36) in this case is just what is predicted by the
hypothesis that ‘nigger’ is semantically equivalent to ‘black’ but that ‘nigger’ is
associated with certain felicity conditions regarding what attitudes are expressed
when the word is used in certain contexts.13
13
A referee pointed out that people don’t refuse to assent to (36) only because it is infelicitous. We
agree. We aren’t trying to give the full explanation of the strength of people’s refusal to assent to (36).
Presumably the full story would refer to the deplorable attitude one would be conveying by asserting (36),
not merely the infelicity of assenting without having the attitude.
123
A. Sennet, D. Copp
What about thinking rather than asserting? Is drawing (33) as a conclusion racist
if (33) is only thought? How can thinking (33) be racist given that thinking (32) is
not racist, unless (33) and (32) do not express the same proposition? The story about
felicity conditions may not be of any help here since felicity conditions govern the
use of words in utterances, but in thinking (33), we are not uttering anything. There
are difficult questions here about whether we think using a language of thought and
whether the dynamics of thought are similar to the dynamics of conversation. We
think that they plausibly are, in the case of episodic, occurrent thought—if it’s
clearly raining outside, there is something infelicitous about thinking ‘it might be
raining’ even though it’s clearly true that it might be raining. The same doesn’t hold
for the more dispositional notion of thinking; but that’s not the relevant sort of
thinking involved in drawing conclusions. So we think it would be racist to think
(33). It is presumably racist to think that MLK was African American but smart,
while it’s not racist to think that MLK was African American and smart.14
Frege cases provide the strongest prima facie case, in our view, for semantic
theories of pejoratives, but we think the argument from Frege cases should be
resisted. Explaining this will require a bit of careful thought. We will be primarily
concerned with the conditions under which one is in a position to know whether a
sentence is analytic.
To see why this is a crucial issue, consider our epistemic position with respect to
‘kike’ and ‘Jew’. Hom and May compare the following sentences:
(13) Jews are Jews.
(14) Jews are kikes.
They claim that while (13) is analytic, (14) is not. (In fact, they think it is false.)
They conclude that ‘Jew’ and ‘kike’ have different meanings or senses.15 However,
the premise of this argument presupposes that we are in a position to know whether
‘kike’ and ‘Jew’ have the same meaning, for otherwise we would not be in a
position to conclude that (14) is not analytic. We have a great deal of evidence,
however, that at least some of us do not know whether these terms have the same
meaning. There is a voluminous philosophical literature that disagrees heavily on
what words like ‘kike’ mean! The paradox of analysis is lurking here, but we think
that no matter how we resolve the paradox of analysis, we have to leave room for
cases in which a person is competent with a term without knowing whether it means
the same as some other term. In the case at hand, the debate concerns the right
14
We owe this idea to Jason Stanley, who floated the idea that thinking and conversing appear to have
analogous dynamics. There are difficult issues here, however, that go beyond the scope of this paper. We
hope the abuse of use and mention in this paragraph is justified by making the text more readable.
15
Hom and May (2014) argue that a non-racist who rejects anti-Semitism cannot consistently accept that
‘kike’ and ‘Jew’ have different senses and also hold that (14) is true. They recommend on this basis
denying (14). In effect, they are here arguing that a semantic theory of the pejoratives must be a DE
theory. We do not understand their argument so we set it aside.
123
What kind of a mistake is it to use a slur?
16
Hom and May (2014) object that on this kind of view, a non-racist would have to eschew the use of
‘kike’ in both speech and thought and that therefore she would be unable to formulate (14). She would be
unable to form the thought that Jews and kikes are one and the same. On this kind of view, they contend,
the thought that Jews are kikes could only be formulated by someone in the state of mind of an anti-
Semite. But their argument rests on a misunderstanding of the kind of non-semantic theory we are
describing. According to the kind of view we are describing, it is infelicitous to use ‘kike’ either in speech
or thought unless one has a relevant negative attitude toward Jews. This is the kind of infelicity involved
in addressing a person using the French ‘tu’ if one is not on familiar terms with that person. But nothing
prevents a Frenchman who is not on familiar terms with you from using ‘tu’ infelicitously in speaking or
thinking of you. Similarly, on the kind of view we are describing, nothing prevents a non-racist from
formulating (14) even though it would be infelicitous of her to do so. As a non-racist who is linguistically
competent she would eschew the use of ‘kike’ in speech and thought but this would not make her unable
to formulate (14). It would only make her uncomfortable to do so.
123
A. Sennet, D. Copp
6 Lewisian charity
17
See Sider (2001) for a description of Lewis’ notion of ‘best fit’ semantics.
123
What kind of a mistake is it to use a slur?
DE theory according to which the extension of ‘kike’ differs from that of ‘Jew’
despite the fact that such a theory also does not fit with the usage dispositions.
(c) We could claim that the overall pattern of use makes the set of Jews the best fit
for the extension of ‘kike’, thereby rejecting semantic DE views, but leaving it open
whether some SE theory is correct or whether instead a non-semantic theory is
correct. Or (d) we could claim that the meaning of ‘kike’ is indeterminate because
there is no best fit. We claim that option (c) is best on Lewisian grounds.
This claim does not force us to disagree with Hom and May’s implicit view that
competent speakers of English believe (or should believe) it is analytic that kikes
are people who ought to be treated badly because they are Jewish. For we could
maintain, given (c), that the case of ‘kike’ is analogous to the case of ‘contact’. The
putative intuition that it is analytic that kikes are people who ought to be treated
badly because they are Jewish is downgraded to a false belief. The overall pattern of
use may militate against respecting this intuition.
We think that (c) is the right approach. But as long as (c) is at least plausible, we
score a sizeable victory because the plausibility of (c) implies the plausibility of
rejecting semantic DE theories. Beyond this—given natural assumptions—given the
plausibility of (c), it is unclear whether ‘Jews are kikes’ is analytic. This, in turn,
undermines Hom and May’s argument in 5.4 about Frege Cases. For consider.
(13) Jews are Jews.
(14) Jews are kikes.
Hom and May claim that (13) is analytic but (14) is not, and they claim that this
supports their view. We respond with the following argument:
P1. If (a), (b), and (c) are all plausible theoretical options, we ought to be agnostic
as to whether the meaning of ‘kike’ differs from the meaning of ‘Jew’.
P2. If we ought to be agnostic as to whether the meaning of ‘kike’ differs from the
meaning of ‘Jew’, then we ought to be agnostic as to whether ‘Jews are kikes’ is
analytic.
C. If we ought to be agnostic as to whether ‘Jews are kikes’ is analytic, then we
don’t know whether ‘Jews are kikes’ is analytic.
Given that (c) is plausible, then even if (a) and (b) are also plausible, we can
conclude that we don’t know whether ‘Jews are kikes’ is analytic. Consequently, a
non-semantic view of the pejorative content of ‘kike’ is not defeated merely on the
basis of the putative intuition that ‘Jews are kikes’ is not analytic.
We think it is fairly clear that intuition cuts against Hom and May’s view. We
will explore the counter-intuitive implications of their view and of other DE theories
in the following section of the paper. Before we do so, however, we shouldn’t
neglect the possibility that option (d) is correct. Perhaps it is indeterminate whether
‘kike’ has the same meaning as ‘Jew’, and perhaps there is no fact of the matter as to
what it means.18 We find this view to be unpalatable although we aren’t sure how to
argue against it. To argue in its favor, however, one would need to argue against
18
Sider (2001) argues for an analogous position regarding criteria of personal identity.
123
A. Sennet, D. Copp
option (c), and we have contended that, so far, the arguments against (c) are less
than convincing. We do think there are good arguments against options (a) and (b),
as we now proceed to show.
We will spend some time in this section arguing against Hom and May’s view, and
so far as we can tell, our arguments generalize to the other semantic theories we are
aware of. Our arguments may not tell against any and all semantic theories, but they
do tell, so far as know, against every such account in the current marketplace of
ideas.
Hom and May’s theory has counter-intuitive implications regarding the following
claims.
(37) All kikes are Mormons.
(38) No-one deserves negative moral evaluation in virtue of being Jewish, not
even kikes.
(39) Kikes deserve negative moral evaluation, but not because they are Jewish.
First, on Hom’s and May’s view, (37) is a necessary truth.19 This is because (37)
is equivalent to (40).
(40) If anyone is a kike, he is Mormon.
Moreover, (40) is necessarily true on Hom and May’s view because the antecedent
of (40) is necessarily false. The antecedent of (40) is necessarily false on their view
because it is necessarily false that anyone deserves negative moral evaluation
because of being Jewish, and so it is necessarily false that anyone is a kike. It
follows that (40) is necessarily true, and hence, (37) is necessarily true. Intuitively,
however, (37) and (40) are both false. They are false given the contingent fact that
many people are Jewish without being Mormon.
Second, on Hom and May’s theory, (38) is self-contradictory because it is
analytic on their account that kike’s deserve negative moral evaluation on account
of being Jewish. (38) is equivalent, on their account, to
(41) No-one deserves negative moral evaluation in virtue of being Jewish, not
even those who deserve negative moral evaluation in virtue of being Jewish.
But (41) is analytically false. Since their account treats (38) as equivalent to (41),
and since (41) is self-contradictory, their account implies that (38) is also self-
contradictory. This seems highly implausible to us.
19
Anderson and Lepore (2013b, p. 361, fn. 20) make a similar point.
123
What kind of a mistake is it to use a slur?
123
A. Sennet, D. Copp
The null extension thesis gets Hom and May into additional difficulties. They need
to explain why it is intuitively true that
(42) All kikes are Jews.
Their explanation turns on their thesis that it is necessarily false that anyone is a
kike since it is necessarily false that anyone deserves negative moral evaluation
because of being Jewish. It follows, on their view, that the antecedent of (43) is
necessarily false.
(43) If anyone is a kike, he is a Jew.
It follows that (43) is necessarily true, and, since (42) is equivalent to (43), (42) also
is necessarily true. Hence it is necessarily true that all kikes are Jews.
We have already seen one problem with this explanation. It implies that (37) is
also trivially true.
(37) All kikes are Mormons.
And worse, their explanation implies that many other problematic sentences are
trivially true including
(44) All kikes are ace pilots.
A further problem with their explanation is that it is much more plausible in
general that, in cases where there are no Fs, sentences of the form ‘all Fs are G’
exhibit presupposition failure.20 Consider:
(45) All weightless elephants are ace pilots.
Intuitively, (45) is infelicitous, which is an indication of presupposition failure. But
on this showing, Hom and May’s account should treat (42) as also exhibiting
presupposition failure, and therefore being infelicitous.
Hom and May therefore are in an awkward position. They need to explain why
(42), ‘All kikes are Jews’, seems intuitively to be true even though on their view
there are no kikes. On their official explanation, (42) would have the same truth
value as (37), ‘All kikes are Mormons’, and (44), ‘All kikes are ace pilots’, but
intuitively it does not. Intuitively (42) is true although (37) and (44) are false. So
their official explanation gives the result they want for (42) but it gives the wrong
result for (37) and (44). Moreover, their official explanation ignores that, on their
account, because there are no kikes, (42) plausibly should exhibit presupposition
failure. Their official explanation of the truth of (42) therefore seems to require them
to eschew appeal to the phenomenon of presupposition failure in the case of (43),
(44), and (45). They apparently have to treat (43), (44) and (45) as trivially true if
they treat (42) as trivially true. Alternatively, they might treat these sentences as
20
We are noncommittal as to whether presupposition failure robs sentences of a truth value in the
relevant contexts. It isn’t relevant to our point.
123
What kind of a mistake is it to use a slur?
exhibiting presupposition failure, but then they would apparently have to treat (42)
the same way. Yet, intuitively, (42) is true. Intuitively, (42) is offensive and
derogatory but it is not infelicitous in the way that it would be if it were a case of
presupposition failure.
7.3 The theory does not account for debates about how to treat people
We can imagine anti-Semites debating how the people they refer to as ‘kikes’ ought
to be treated. Some think that they ought to be driven from the community. Others
think they ought to be pitied because their ancestors murdered Jesus. Another says
they ought to be feared because they aim to take over the world. One says that at
least they ought to be the target of negative moral evaluation simply because they
are Jewish. Someone responds that that is foolish. They ought to be the target of
negative moral evaluation because they are exploiters, not simply because they are
Jews. And so on, for all the crazy and frightening ideas you can imagine anti-
Semites entertaining.
Now it does not seem to us that the claim that ‘kikes’ ought to be the target of
negative moral evaluation because they are Jewish has a special status in such a
debate in that it cannot coherently be denied. An anti-Semite could deny it. An anti-
Semite might say, for example, that negative moral evaluation of all ‘kikes’ is not
appropriate because at least the newborn ‘kikes’ do not deserve negative moral
evaluation. Like all ‘kikes’, he might say, they deserve contempt, not negative
moral evaluation. The debate might be nutty, but not because it is incoherent to
deny that negative moral evaluation would be appropriate.
The point we are making here is related to our earlier point that (39) is not self-
contradictory. Our imagined debate among anti-Semites is intended to draw out the
intuition that neither (39) nor (46) are self-contradictory. They rather make
substantive claims that an anti-Semite could coherently accept:
(39) Kike’s deserve negative moral evaluation, but not because they are Jewish.
(It’s because they are greedy.)
(46) Not all kikes ought to be the target of negative moral evaluation since the
newborn kikes do not yet deserve negative moral evaluation.
The objection we have been making so far is again a specific objection against
Hom and May’s account. The more general objection arises from the fact that a
semantic theory aims to explain the offensive character of a pejorative by supposing
that the meaning of the pejorative includes an offensive component, such as an anti-
Semitic component in the meaning of ‘kike’. The problem is that bigots can disagree
as to why the objects of their bigotry are contemptible and they can disagree
whether contempt or negative moral evaluation or some other attitude is
appropriate. A semantic theory, whether a DE theory or an SE theory, must make
a choice about this and build some thesis about the appropriate attitude and the basis
for its being appropriate into the meaning of the pejorative. They will therefore most
likely be committed to treating as self-contradictory certain claims, such as (39) or
(46), which a bigot could coherently accept. We therefore think that the objection
123
A. Sennet, D. Copp
we have been discussing will most likely generalize and yield an objection to any
semantic theory, whether a DE theory or an SE theory.
Bigots with respect to different groups of people can have different views about
the attitudes that are appropriate toward those groups and about why the attitudes
are appropriate. This means that a semantic theory is at risk of over-simplifying
in such a way that it does not plausibly generalize to all pejoratives. Hom and
May’s theory illustrates the problem. We have already mentioned the examples
of ‘retard’ and ‘dwarf’. Similarly, ‘redneck’ is a pejorative. But those who use
the term generally don’t think rednecks ought to be the target of negative moral
evaluation because they are members of the white rural working class. They
have contempt for members of the white rural working class, but needn’t think
there is any moral fault that is shared by all such people simply because they are
members of the white rural working class. Another example is the term
‘honkey’. This term is a pejorative, and yet those who call white people
‘honkeys’ would not plausibly be thought committed to the idea that all white
people are the appropriate targets of negative moral evaluation just on account of
their being white.
123
What kind of a mistake is it to use a slur?
123
A. Sennet, D. Copp
8 Conclusion
123
What kind of a mistake is it to use a slur?
We have been using Hom and May’s (2013) account to illustrate our objections
to semantic theories of pejoratives because we suspect that our criticisms generalize
to any other semantic view, whether a DE theory or an SE theory. Certainly we
think that they generalize to Hom’s (2008) view. We don’t know of any semantics
views to which they fail to generalize.
Our argument has been complex, and it would be difficult to summarize it briefly.
We begin by mentioning an argument that is intended specifically as an argument
against DE theories—theories according to which a pejorative has a different
extension from its neutral counterpart. We contend that views of this kind are
undermined by the fact that pejoratives are used to refer to the same people as the
corresponding neutral counterparts are used to refer to. ‘Kike’ is used by anti-
Semites to refer to Jews and everyone who knows how ‘kike’ is used knows this.
There is no uncertainty in the minds of those who are competent with the use of
‘kike’ regarding to whom an anti-Semite means to refer when using ‘kike’. This
point undermines DE theories. And against SE theories, note that semantic theories
all say that a pejorative and its neutral counterpart make different contributions to
the truth-conditional content of the sentences in which they appear. Given this, it is
difficult to think of a plausible SE semantic theory since such a theory would have
to say that a pejorative and its neutral counterpart make different contributions to
the content of the sentences in which they appear even though there is no difference
in their extension. It is difficult to see what the difference could be if not one that
implies a difference in the extension of the terms. Moreover, finally, any semantic
theory, whether a DE theory or an SE theory, will imply certain sentences to be
analytic that might be denied or at least debated by bigots. All of these
considerations weaken the plausibility of semantic theories, or so we argue. Our
main argument against semantic theories, however, is that they fail to explain the
offensiveness of pejoratives. They do not explain why it is belittling and demeaning
to use a pejorative in referring to a person or group but not to use its neutral
counterpart.
Consider, for instance, the DE theory we mentioned before according to which
PEJ(N) means ‘N that is contemptible for being N’. Like Hom and May’s view, this
account also implies the truth of (59), ‘No chink will ever teach at this university’,
which it treats as semantically equivalent to
(61) No-one will ever teach at this university who is of Chinese ancestry and who
deserves contempt simply because he is of Chinese ancestry.
Obviously, however, (61) is something any right thinking person could endorse
because no-one deserves contempt because of his ancestry, and so no-one who will
ever teach at the university deserves contempt because of his ancestry. But whereas
(61) is something a right thinking person could endorse, (59) is not, because (59) is
derogatory and pejorative.
Or consider the SE theory according to which PEJ(N) means ‘N for whom anti-
Ns would have contempt’. On this view, (59) is semantically equivalent to
(62) No person of Chinese ancestry for whom people who are anti-Chinese would
have contempt will ever teach at this university.
123
A. Sennet, D. Copp
But whereas (59) is both derogatory and pejorative, (62) is neither, and a person
who asserts it may well hope it is false. Hence, again, these semantic accounts do
not explain the nature of derogatory and pejorative language. We therefore
recommend looking elsewhere for a theory of pejoratives.
References
Anderson, L., & Lepore, E. (2013a). Slurring words. Noûs, 47(1), 25–48.
Anderson, L., & Lepore, E. (2013b). What did you call me? Slurs as prohibited words. Analytic
Philosophy, 54, 350–363.
Bach, K. (1999). The myth of conventional implicature. Linguistics and Philosophy, 22(4), 327–366.
Bach, K. (2014). Loaded words: On the semantics and pragmatics of slurs. Presented to the 2014
meetings of the Pacific Division of the American Philosophical Association. http://online.sfsu.edu/
kbach/Bach.LoadedWords.APA.pdf.
Camp, E. (2013). Slurring perspectives. Analytic Philosophy, 54, 330–349.
Copp, D. (2001). Realist-expressivism: A neglected option for moral realism. Social Philosophy and
Policy, 18, 1–43. (Reprinted in D. Copp, Morality in a natural world (pp. 153–202). Cambridge:
Cambridge University Press, 2007).
Copp, D. (2009). Realist-expressivism and conventional implicature. Oxford Studies in Meta-Ethics, 4,
167–202.
Croom, A. (2013). How to do things with slurs: Studies in the way of derogatory words. Language &
Communication, 33, 177–204.
Hom, C. (2008). The semantics of racial epithets. Journal of Philosophy, 105, 416–440.
Hom, C. (2010). Pejoratives. Philosophy Compass, 2, 64–185.
Hom, C., & May, R. (2013). Moral and semantic innocence. Analytic Philosophy, 54, 293–313.
Hom, C., & May, R. (2014). The inconsistency of the identity thesis. Protosociology, 31 Language and
Value.
Jesion, R. (2013). Slurs and stereotypes. Analytic Philosophy, 54, 239–314.
Kaplan, D. (unpublished). The meaning of ‘oops’ and ‘ouch’.
Lewis, D. (1983). New work for a theory of universals. Australasian Journal of Philosophy, 61, 343–377.
(Reprinted in D. Lewis, Papers in metaphysics and epistemology (pp. 8–55). Cambridge: Cambridge
University Press, 1999).
Lewis, D. (1984). Putnam’s paradox. Australasian Journal of Philosophy, 62, 221–236. (Reprinted in D.
Lewis, Papers in metaphysics and epistemology (pp. 56–77). Cambridge: Cambridge University
Press, 1999).
Miščević, N. (2011). Slurs & thick concepts—Is the new expressivism tenable? Croatian Journal of
Philosophy, 11(2), 159–182.
Potts, C. (2005). The logic of conventional implicatures. New York: Oxford University Press.
Richard, M. (2008). When truth gives out. Oxford: Oxford University Press.
Sider, T. (2001). Criteria of personal identity and the limits of conceptual analysis. Philosophical
Perspectives, 15, 189–209.
Whiting, D. (2013). It’s not what you said, it’s the way you said it: Slurs and conventional implicature.
Analytic Philosophy, 54, 364–377.
123