Section 2.3.2, sub II, has shown that er-nominalization of intransitive PO-verbs is less common than that of regular intransitive verbs. It will therefore not be a big surprise that the primed examples in (390) show that the PO-verbs in (388) do not allow er-nominalization. Note, however, that this may not be entirely due to the presence of the PP-complement, since the primeless examples of (390) show that er-nominalizations of the corresponding regular transitive verbs are often unacceptable as well.

A contrast arises only in the case of the (b)-examples: the nominalization can only receive the total affectedness reading of the transitive construction Jan eet een boterhamJan is eating a sandwich, not the part-of reading found in the PO-verb construction Jan eet van de boterhamJan is eating from the sandwich. The contrast between etento eat and the other input verbs in terms of er-nominalization may reflect the fact that the transitive use of eten is also more common.

Like the intransitive PO-verbs discussed in Section 2.3.2, sub II, the PO-verbs in (388) all take the auxiliary hebben, which is consistent with the assumption of an external argument; the examples in (391) show that these verbs behave in this respect like the corresponding transitive verbs.

The past/passive participles of the PO-verbs in (388) cannot be used attributively to modify the subject of the corresponding verbal construction, whereas their present participles can. In this respect they behave like the intransitive PO-verbs discussed in Section 2.3.2, sub II. The ungrammaticality of the examples in (392) with a past/passive participle is compatible with the assumption that the PO-verbs in (388) take an external argument.

All PO-verbs in (388) allow passivization, which is sufficient to assume that these verbs take an external argument. The last two examples may feel somewhat awkward, but examples of this kind do occur on the internet.

The data in the previous subsections shows that the PO-verbs in (388) behave in all relevant respects like the intransitive PO-verbs discussed in Section 2.3.2, sub II. Therefore, apart from the fact that the PP-complements of these verbs alternate with nominal complements, nothing special needs to be said about these verbs.

Although the transitive PO-verbs in (394) take an agentive subject, er-nominalization seems to produce a marginal result; in this respect they behave like (most of) their intransitive counterparts in Subsection I.

The transitive PO-verbs in (394) select the auxiliary verb hebbento have, just like their ditransitive counterparts. This is consistent with the assumption that these verbs take an external argument.

Past/passive participles of the PO-verbs in (394) cannot be used attributively with a noun corresponding to the nominative argument of the corresponding verbal construction. However, it seems marginally possible to use them when the modified noun corresponds to the dative object; cf. Section 2.1.3, sub IID. This is also the case when the PO-object is replaced by a direct object, although some speakers seem to like this option (even) less; this is shown in (398).

Replacing the past participles in (398) with a present participle triggers an agentive reading on the modified noun.

The PO-verbs in (394) allow passivization. The assumption that the nominal complements are datives is motivated by the fact that it is often claimed that they cannot be promoted to subject; passivization is taken to result in the impersonal passive in the primeless examples in (400). Note, however, that some speakers at least marginally allow the nominal complement to become the subject of the clause with the PO-verbs in (394). For such speakers, the primed examples are also more or less acceptable, and cases like ... worden om advies/raad/hulp/gevraagd ... be asked for advice/council/help can be found with all subject pronouns on the internet.

The passivization test seems to be a good tool to distinguish PO-verbs with a dative object from the transitive PO-verbs discussed in Section 2.3.2. The latter do not allow impersonal passivization, as can be seen from the contrast between the two (b)-examples in (401); the impersonal passive in (401b) is excluded, so that the nominal object must be promoted to subject.

However, the passivization test is not always easy to use. For example, normative grammarians have claimed that the PO-verb wijzen op in (402a) takes an indirect object, and that the passive construction in (402b) is consequently an impersonal passive; the noun phrase does not function as a subject, and the finite verb should therefore have (default) singular agreement; cf. taaladvies.net/taal/advies/vraag/917. This claim goes against our intuitions, according to which example (402b) is only possible with plural agreement (i.e. the form actually found in speech); cf. onzetaal.nl/taaladvies/advies/de-luisteraars-werd-werden-erop-gewezen. This strongly suggests that (in contemporary Dutch) we are not dealing with an impersonal but with a regular passive, which is confirmed by the fact that using the object form (favored by normative grammarians) of the pronoun in (402c) leads to a severely degraded result. We think that this conclusively shows that the normative grammar is wrong and that we are not dealing with a PO-verb with a dative object but with a transitive PO-verb, which is perhaps further supported by the fact that the German translation of wijzen op (hindeuten/hinweisen auf) also takes an accusative object.

For completeness’ sake, note that ditransitive verbs with an indirect object and a clausal direct object often have a similar choice between impersonal and regular passivization, as illustrated by (403); some speakers allow a dative object to be promoted to subject when no accusative object is present (e.g. because there is an object clause); cf. Section 3.2.1.3, sub II, for further discussion.

For our present purposes the contrast between the types of passivization is not very important: the mere fact that the PO-verbs under discussion allow (impersonal) passivization is sufficient to conclude that they take an external argument.

Dative arguments usually precede PP-complements; in non-contrastive contexts, the latter can only precede dative objects if they are moved into clause-initial position by wh-movement (not shown here).

The data in the previous subsections shows that the PO-verbs in (392) have an external argument, which is especially clear from the fact that they allow passivization. Of course, the passivization is different from that found in the corresponding transitive constructions, since the internal nominal argument is dative, not accusative. As a result, it is usually the impersonal passive that is found, although some speakers also allow passive constructions in which the dative argument of the active construction is promoted to subject. This does not detract from our main finding that impersonal passivization is the crucial diagnostic that distinguishes PO-verbs with a dative nominal argument from those with an accusative one.