Why some verbs can form a resultative construction while others cannot: Decomposing semantic binding

In this paper it is argued that the occurrence of adjectival resultative constructions in German is grounded on the semantic properties of the verb. The predicting properties can be subsumed under the umbrella concept of semantic binding (SB), which in the case of resultative constructions is the semantic binding of a resultant state (SB-R). It is hypothesized that SB-R is a complex concept which excludes specific verb classes from having a resultant construction and which distributes over a gradual scale of verb classes which vary in the way they can bind additional semantic elements. A regression analysis was applied to a set of seventy-seven German verbs, the criterion variable being SB-R values, and a series of semantic properties as predictor variables. The analysis yielded a model that accurately predicts the SB-R value of each verb and establishes a distinction in six verb classes. The regression model is elaborated in linguistic terms, on the basis of three elementary semantic properties: (1) the time scheme of the verb, (2) the affectedness of the grammatical subject and (3) the affectedness of the object, properties that also play a role in defining transitivity, which can be seen as the SB of an object. The resulting model provides a transparent algorithm for determining the occurrence of adjectival resultative constructions in German.