There is a lot of debate about the meaning of slurs, which are words that refer to a certain group of people in a derogative manner. Many scholars (e.g. McCready 2010; Croom 2011; Gutzmann 2015) consider slurs, unlike their ‘neutral’ counterparts, to have both descriptive and expressive content. In this squib, I discuss initial data on the use of Dutch slurs and related terms on Twitter. The Dutch data show that ‘neutral’ labels sometimes seem to behave like slurs to some degree and that some slurs can be used as swear words without (seemingly) conveying any descriptive property. Based on these observations, I suggest that slurs may be part of a continuum of labels, ranging in the degree to which they have descriptive versus expressive content, entailing a gradual distinction between conversational and conventional implicatures in the lexical domain. A gradual view on conventionality in the lexical domain would moreover pave the way for a less fundamental distinction between expressive and social meaning. It is concluded that the debate on slurs would benefit from a broader and more empirical approach to pejorative terms for (groups of) people, as their actual use and interpretation can show a mixture of (degrees of) descriptive, expressive and social meaning.