This section has discussed the dative alternation and has shown that there are at least five semantic subtypes, which can be syntactically distinguished by means of the form of the periphrastic indirect object:

      We have seen that there are reasons for assuming that periphrastic indirect objects are not internal arguments (that is, PP-complements) of the verbs: the aan-, naar- and van-PPs in (442a-c) clearly behave as complementives. Possessive bij-phrases have been shown not to be selected by the verb at all but to form a constituent with the locational phrase that contains the possessee. Something similar may hold for benefactive datives/voor-PPs; the fact that they are optional and not semantically implied by the verb suggests that they are not arguments of the verb, but adverbial modifiers.
      Our investigation has further shown that there are reasons for assuming that at least the double object constructions in (442a-c) are derived from an underlying structure that is very similar to that of the periphrastic construction, and we may therefore conclude that the double objects in these constructions do not function as internal arguments of the verb either. Although we did not discuss this here, Broekhuis & Cornips (1997) have argued that the possessive dative and the possessive bij-phrase likewise share a common underlying form: if so, this implies that possessive datives do not function as internal arguments of the verb either.
      We refer the reader to Hoekstra's (2004) Small Clauses Everywhere for a discussion of the theoretical ramifications of the conclusion that indirect objects do not function as internal arguments of the verb, and to Den Dikken (1995:ch.3) for a theoretical account of the dative alternation that to our minds seems to fit best the data described in this section as well as for a critical review of a number of alternative proposals.