Adjectives and degree modification conclusion

This chapter started out with the question of how adjectives and gradability are related to one another, and I have looked at this question from the perspective of the distribution of degree expressions. It has been argued that the categories degree expressions may combine with form a continuum in the sense that degree expressions cannot combine with two categories that are not adjacent on this continuum, unless they may be combined with all categories in between as well. Some degree modifiers can be used in all contexts (e.g., trop) while others are only compatible with adjectives (e.g., too) or with plural nouns (e.g., many). In between these extreme cases all sorts of “intermediate “distributions are found.

In the second part of the chapter I explored several aspects of the continuum in relation to the status of adjectives. First I explored different hypotheses concerning the presence vs. absence of degree variables. A first possibility is to assume that gradable adjectives contain a degree variable, while other gradable categories do not. This allows us to assume that type A expressions have a semantics that makes them incompatible with expressions that do not offer a degree variable. This approach would force us to postulate systematic ambiguity for type C expressions (the expressions that can be used with all categories). It has been shown that this type of approach does not account for syntactic differences between type A and type C expressions in the context of adjectives. A second possibility is to assume that all contexts that are compatible with type B expressions (gradable adjectival, nominal, and verbal predicates) are sensitive to the presence of a degree variable. This would have the (possibly) desirable result that abstract gradable verbs such as to appreciate and gradable adjectives obtain their degree reading in a similar way. Finally, one could assume that all degree expressions are sensitive to the presence of a degree variable that would then be available cross-categorially (see Cresswell 1977). This would allow us to do away with the ambiguity of expressions such as more and less. In the latter two scenarios (degree variable in all gradable predicates and degree variable everywhere), the special behavior of adjectives remains unexplained.

A second way to look at the distribution of type A expressions involves scale structure. Building on Kennedy andMcNally (2005), it has been hypothesized that certain types of scales are typically adjectival. This type of approach is actually compatible with different views on the presence versus absence of a degree variable. On the one hand scale structure might explain certain distributional patterns within a Cresswellian approach to the degree variable. On the other hand, under the assumption that only adjectives contain degree variables, the special scalar properties of adjectives might be related to the presence of this degree variable.