‘Metalinguistic’ negations, denial, and idioms

A number of unusual negations are typically marked in various ways and have irregular interpretations. I seek to determine what these negations mean or convey, what exactly they deny, and how their irregular interpretation is related to their regular interpretation. I argue that the “marks” Horn identified are not semantically significant, and that the common term ‘metalinguistic negation’ is a misnomer. Contrary to Horn, Burton-Roberts, and Van der Sandt, irregular negations are not always used to object to previous utterances. I make the case that what an irregular negation denies is an implicature of its root. Different types of irregular negation deny different root implicatures. Irregular negations exist because of three levels of linguistic conventions. While one type of irregular negation involves live implicature, I defend the hypothesis that the others are idioms (partially non-compositional compounds) that plausibly arose when implicatures died. I extend my account of how irregular negations can be so productive despite being idioms by generalizing the scalar implicature denial convention and identifying the relevant idiomatic forms for other irregular negations.