The authors embrace option (C), while keeping the idea, expressed in option (A), that group membership is not a morally evaluable feature for them: G is a group that is/was the target of unjust, hateful or discriminatory ideology.
I agree with Hom and May about rejecting option (B) as a non-starter, but I do not embrace their solution. On my account any property can in principle individuate a target group, as long as there are speakers who find it interesting or convenient to use a piece of language to pick out such a property while conveying an evaluation of the objects that instantiate it. Note that my view imposes no requirement on whether the evaluation is warranted. My criterion to define a slur differs from Hom and May’s (and from most accounts of slurs) in that (i) slurs do not have to be wrong by definition, as it is not part of how I define a slur that the triggered evaluation is wrong and “unjust,”2 (ii) for a slur to exist, it is not necessary to have a culturally ingrained discriminatory ideology within society—it suffices that a group of speakers coin a term to express a systematic evaluation of people instantiating a certain property, and (iii) even though prototypical examples of slurs convey a negative evaluation, this is not an essential feature; it is conceivable that a slur might convey a positive evaluation. An example of what a slur with a positive polarity may look like—and again, having a positive polarity does not mean being just—is ‘Aryan’: Nazis used this term for Indo-Europeans, while conveying some positive evaluation, thus supporting the idea that being Indo-European is good in and of itself.
Slurs turn out to be very similar to thick terms in this framework, where the polarity of the evaluation could be any in principle (even though it is typically negative), and where the target class of a slur could be any (even though it usually concerns ethnicity, sexual orientation, and the like). Most slurs are both negative and unjust—they convey a negative evaluation of their targets on the basis of characteristics that in no way ground or justify derogation, exclusion, and so on. However, the linguistic mechanism through which slurs encode values does not need their evaluative content to be negative or unjust.
1.1.2 Distinguishing Thick Terms from Thin Terms
A further question—in addition to how to distinguish HEs from ordinary descriptive terms used evaluatively—is how to characterize various kinds of evaluatives, for instance, how to discern hybrid and pure evaluatives. In other words, how can one draw the divide between thick terms, such as ‘generous’ or ‘lewd,’ and thin terms, such as ‘good’ and ‘bad’? The thin versus thick distinction, formulated in metaethics (Williams 1985), is far from being uncontroversial, even though it is widely accepted. For Eklund (2011: 40), if one analyzes the evaluative content of thick terms as conventional implicatures (or presuppositions, I would add), then one ends up drawing a “dramatic difference” between thin and thick terms, since for the former the evaluation belongs to the Fregean sense, whereas for thick terms it would belong to the Fregean tone. Väyrynen (2013), who defends the view that thick terms do not encode evaluations, discusses some of the major difficulties raised by such a distinction. According to Väyrynen, the evaluative content associated with thick terms arises from conversational mechanisms that can vary on a case-by-case basis. In Cepollaro and Stojanovic (2016), we defend the opposing view that HE-evaluations are recorded in the meaning of such expressions as evaluative presuppositions. Thus our view has to answer the question of what distinguishes pure evaluatives from hybrid ones.
One way to distinguish thin and thick terms is to rely on the notion of at-issueness: the evaluative content is at-issue for thin terms and presupposed—hence not-at-issue—for HEs, whose descriptive content is instead what is at issue. The at-issue versus not-at-issue distinction, widely accepted in discourse analysis,3 amounts to the idea that utterances typically have something like a main point, i.e. a so-called question under discussion (QUD); the at-issue propositions associated with an utterance are relevant to the QUD, in the sense that they entail “a partial or complete answer to the QUD” (Simons et al. 2010). Every proposition associated with an utterance that is not relevant to the QUD is labeled as ‘not-at-issue.’ Presuppositions are typically not relevant to the QUD, as they amount to content that is taken for granted. Going back to the thin versus thick distinction, the idea is that the evaluation is not-at-issue (it is taken for granted) in the case of thick terms, but at-issue (it is the main point, relevant to the QUD) in the case of thin terms. Although this idea sounds plausible to a certain extent, the picture appears to be quite complicated, as the HE-evaluation can become precisely what is at issue:
1 6.A.Madonna’s show was lewd.B.I disagree. It was sexually explicit beyond conventional boundaries, indeed, but it was not bad in any way.
However, this is not surprising after all: as I will discuss in greater detail in chapter 3, presupposed content can become at-issue under the appropriate circumstances; the cases of negotiation involving metalinguistic disagreement—like (6)—are particularly interesting for evaluatives. For the time being, I consider other examples that seem to challenge the idea of a sharp divide between thick and thin terms.
The first problems are raised by those terms that are thought to be on the edge between thin and thick, such as ‘just.’ ‘Just’ seems to have some descriptive content, but very little so to say. This impression of being on the edge between the thick and the thin can be explained in terms of how specific the descriptive content is. For example, the descriptive at-issue content of ‘just’ is quite vague, something like ‘in accordance with standards and requirements.’ In contrast, most thick terms are more specific than that in their at-issue content (and slurs even more so): for instance, ‘lewd,’ meaning ‘sexually explicit beyond conventional boundaries,’ and ‘wop,’ meaning ‘Italian.’ However, one could simply acknowledge that some thick terms have a very general descriptive content that needs to be contextually specified, without giving up the distinction between thin and thick terms.
In Cepollaro and Stojanovic (2016), we argued that further difficulties for postulating a sharp boundary between the thin and the thick stem from those expressions that are ‘thicker’ than the all-purpose evaluatives ‘good’ and ‘bad’ and nevertheless resemble very closely such expressions in that they (i) are typically used to convey evaluative content and (ii) do not have a rich descriptive content in addition to the evaluative one. We considered two such cases. Highly positive and highly negative expressions, such as ‘awesome,’ ‘excellent,’ ‘fantastic,’ ‘magnificent,’ on the one hand, and terms like ‘awful,’ ‘horrible,’ ‘terrible,’ on the other hand. The above terms look like intensified thin terms (something like ‘good (or bad) to a very high degree’). However, there seems to be some additional lexical information that distinguishes those expressions from each other, that is, some descriptive meaning: ‘excellent’ is for what is superior to the rest in a certain area, ‘terrible’ is for what provokes terror and so on. The second case includes terms that look like thin terms, but in relation to a specific domain: ‘beautiful,’ ‘ugly,’ or ‘evil.’ As a matter of fact, when discussed within the realm of aesthetics, ‘beautiful’ and ‘ugly’ are taken to be thin terms. Similarly, ‘evil’ would likely count as thin in ethics, as it means something along the lines of ‘morally bad.’ On the other hand, if instead of thinking about evaluatives from a specific domain, one considers them from a broader perspective, ‘beautiful’ and ‘ugly’ are narrower than the all-purpose ‘good’ and ‘bad,’ because the former constrain the evaluation to the aesthetic field; similarly, ‘evil’ is narrower than ‘bad,’ as the negative evaluation that it encodes is linguistically constrained to the moral domain. We drew the conclusion that terms like ‘excellent’ and ‘horrible,’ ‘beautiful’ and ‘ugly’ are thicker than ‘good’ and ‘bad,’ but thinner than ‘courageous’ or ‘cruel.’
The problematic cases presented in Cepollaro and Stojanovic (2016) do not inevitably knock down the idea of a distinction between thin and thick terms. Let me sketch a tentative alternative explanation. One could analyze ‘beautiful’ and ‘ugly’ as thin terms restricted