Manner modification of state verbs conclusion

Applying the Davidsonian analysis to state verbs is appealing in its uniformity, as adverbial modification can, in general, be treated as predication over underlying eventualities (in Bach’s (1986) terminology). Both Landman (2000) and Mittwoch (2005) attempt to bolster Parson’s (1990) argument that the semantics of state sentences must involve quantification over underlying state variables by showing that manner adverbial modification of state verbs has properties that require such an analysis. We have argued that this attempt is ultimately unsuccessful and that, in fact, manner modification of state verbs is a highly constrained phenomenon which receives a better analysis as classical predicate modification.

There was much clearing of the brush to be done in distinguishing the relevant set of verbs and adverbs, thus eliminating a large class of apparently problematic cases. We showed that the few types of adverbs that can be used to modify state verbs typically don’t have the semantic characteristics associated with eventuality predication: they don’t have the properties Drop, Permutation, and Non-Entailment. For particular examples that have been raised in the literature which appear to exhibit these properties, we have shown that the observations are either mistaken or best otherwise analyzed.

Manner adverbials that modify state verbs appear to be much more closely associated with the verb they modify than are manner adverbials modifying event verbs (far more closely than we might expect on a Davidsonian analysis). We have made some tentative proposals accounting for this, suggesting that many verb–adverb combinations are a sort of idiomatic expression or collocation. In passing we discussed the category-neutrality of collocations, pointing out that adverb/adjective parallels cannot really be taken as evidence for a Davidsonian account. We have seen that many manner adverbials acquire a degree modifier reading when used to modify state verbs, and attempted to suggest why this might be. We also noted that the fine structure of degree modification predicts precisely which degree modifiers can be dropped preserving truth and which not. Finally we have suggested that there are event-related modification meanings associated with state verb modification and showed how the meanings of these might be derived from the conceptual structure of generic predicates.

While we have left many threads untied, we hope to have made clear that the case for treating “manner” modifiers of state verbs as predicates of underlying states and not as classical predicate modifiers is a weak one, based on a superficial similarity. While it may still be the case that such an analysis is defensible, there are a number of issues to be addressed. There is the restriction of manner modification of state verbs to gradable predicates, the fact that manner modifiers of state verbs are so frequently reanalyzed as degree modifiers, and the close lexical selectivity of manner modifiers of state verbs. Each of these needs an explanation, even on a Davidsonian account of state verb modification. We have provided at least preliminary accounts here on a classical account which makes no reference to underlying states. Whether reference to underlying states provides any empirical benefit would remain to be seen.