Anja Goldschmidt's PhD thesis contributes to the study of semantic well-formedness in verb-adverb modification. She shows that conceptual knowledge is relevant for understanding sentences with verb-adverb combinations.
Text meaning and knowledge
The meaning of words and sentences is related to our knowledge of the world. Because we know that fruit has a certain taste, we can easily understand a sentence like 'This fruit is very sweet'. At the same time, we also know that 'This fruit is very fast' is nonsense. For understanding words and sentences, we use a large amount of information about the world, based on our experience with and knowledge of the things that words refer to.
Verbs and adverbs
This can not only be seen with nouns such as 'fruit' but also with verbs. For example, verbs such as 'hitting' or 'pushing' express a force exertion through contact. If we want to express that someone exercises little force, this can be done using the adverb 'lightly', but other adverbs can have a similar effect. The sentence 'Lisa hits John playfully on his arm' is generally understood to mean that Lisa has not hit John hard, because we know that in a playful situation it is usually not the intention to hurt anyone. Our knowledge about the world thus influences how we understand this sentence.
Modeling conceptual underpinnings
Goldschmidt proposes an Event Semantic analysis that models these findings based on underlying force vectors, which represent the force generated by the agent. Adverbs, verbs and prepositions are analysed as having specific requirements on force vectors, which also accounts for observations such as the selectional requirements of verbs for certain prepositions or the unacceptability of specific verb-adverb combinations. The underlying force vectors and other relevant meaning components are then represented in a Frame Semantic model which transparently spells out the conceptual underpinnings of semantic well-formedness in verb-adverb modification.