In this talk, I investigate the morpho-semantics of degree constructions such as synthetic comparatives, periphrastic comparatives and equatives. Some of these constructions are known to give rise to presuppositions of evaluativity that do not follow from the traditional semantics assumed for degree constructions:
(1) Athos is taller than Porthos is.
(2) Athos is less tall than Porthos is.
(3) Athos is less short than Porthos is. (Presupposition: Athos/Porthos count as short.)
Rett (2008, 2014) argues that an adjective gets its evaluative interpretation when it is modified by an optional operator called EVAL, that contributes the reference to a standard of comparison. In turn, Rett shows that (3) is obligatorily evaluative because its non-presuppositional reading is precluded by the sentence in (1) and this by virtue of the fact that the two sentences are semantically equivalent, and that the degree expression 'less short' is marked relative to 'taller'. This notion of markedness advocated by Rett is not fully explicated in her proposal but relies on the idea that 'less' and the negative antonym (like 'short') "encodes the force of a negation".
I will show how Rett's notion of markedness can be restated in term of structural complexity under the Syntactic Negation Theory of Antonymy (Heim 2001, 2008; Büring 2007). This approach captures the antonymic relationship that exists between pairs of adjectives and pairs of polar operators by positing a covert negation LITTLE ([little tall] 'short', [-er little] 'less'). I will then propose an LF principle dubbed 'Structural Economy' (adapted from Marty (in preparation)) that imposes limits on syntactic complexity. In addition, I will show how EVAL interferes with PF-processes in the matrix clause. Finally, I will argue that PF and LF viewpoints are both necessary to understand the distribution of (un)-evaluativity.