There are minimally three suitable cases for an analysis involving an intransitive adposition, i.e. where the adpositional element retains its original meaning and behaves semantically and syntactically like a predicative (locational) PP.

In contrast to the intransitive adpositions discussed in Subsection I, verbal particles need not express a spatial meaning. Often, their meaning is more or less aspectual in nature. For example, the particle opup in (329a) transforms the stative verb staanto stand into an activity verb, and wegaway in (329b) changes the stative verb drijvento float into a process verb.

In addition, verbal particles usually make the event telic: the particle verb is inherently bounded in time and results in a new location of the located object. The fact that the particle verb is telic can also be seen from the fact that it takes the auxiliary zijnto be in the perfect tense, while the stative verbs staan and drijven take the auxiliary hebbento have.

Exceptions are the particles door and mee in particle verbs like doorwerkento continue working and meewerkento cooperate/work along, which are atelic and take the auxiliary hebbento have in the perfect tense; cf. Van Hout (1996:96).

The fact that most monadic particle verbs select the auxiliary zijn clearly shows that they are unaccusative verbs. As shown for weg drijven in (331), they also satisfy the other criteria for unaccusativity; the past participle can be used attributively (which is excluded for op staan, due to a general prohibition regarding contraction verbs, discussed in Section ), and passivization is excluded.

Particles of monadic particle verbs can be seen as predicative elements that function as a complementive. The examples in (332) illustrate this by showing that particles, like resultatives, have the ability to introduce an additional argument into the clause that is not selected by the verb. This means that the noun phrase het meisjethe girl in (332b) is semantically licensed by the particle uit, which implies that the latter must be a predicate of some kind.

The hypothesis that particles are complementives is also supported by the fact, illustrated in (333a&b), that both resultatives and particles must be left-adjacent to the clause-final verb(s). Moreover, since a clause can contain at most one complementive, this hypothesis also correctly predicts that examples such as (333c), in which a verbal particle co-occurs with a resultative phrase, are ungrammatical.

There are some putative counterexamples to the claim that verbal particles and complementives are in complementary distribution. Consider (334). The verb leggento put requires a complementive, and so (334b) confirms our hypothesis that the particle neer is a complementive, just like the PP op de tafel in (334a). Note, however, that the PP op de tafel and the particle neer can occur together in constructions such as (334c). It should be noted, however, that the PP in (334c) does not have to precede the clause-final verb, but can also follow it. This shows that the PP does not function as the complementive in this example.

The exact syntactic function of the PP in (334c) is unclear. For example, Den Dikken (1995a) claims that the PP is actually a complement of the particle (cf. Section 35.2.1.2), and that its logical subject, the noun phrase het boek, is raised to the subject position of the particle. Broekhuis (1992), on the other hand, argues that the PP actually has an ambiguous status: sometimes it acts as the complement of the particle, as claimed by Den Dikken, and sometimes as an independent adverbial phrase. We will not digress on this issue here, but conclude by noting that the behavior of the PP in (334c) is part of a more general pattern, which will be discussed in Section 35.2.1.2, sub IIC.

Another possible counterexample of this kind, involving the adjectival complementive groengreen and the particle af, is given in (335). However, since adjectives never undergo extraposition, it cannot be shown that the adjective does not function as a complementive of the verb in the somewhat marked but passable example in (335c); note that this example becomes perfectly acceptable if we replace afverven with the more lexicalized particle verb aflakkento varnish.

Particle verbs are often involved in verb alternations. Example (336a) is a simple predicative construction involving a change of location. The examples in (336b&c) show that the located object de kleren and the reference object de koffer from (336a) can also appear as the direct object of the clause, as evidenced by the corresponding passive constructions in the primed examples. The fact that the reference object from (336a) surfaces as direct object in (336c) unmistakably shows that the adposition in in (336c) is not a postposition but a verbal particle.

So far, attempts to provide a general description of the meaning contribution of verbal particles have not been successful. It can be aspectual in nature, as in (329), or it can express a location, as in (336), but it can also add a more or less systematic meaning aspect to the verb: for example, the particle door in the primed examples in (337) changes the meaning of geven from “to give” to something like “to pass”. Note that the we find the same change with a periphrastic indirect object with aanto; cf. Jan gaf het boek aan Marie (door) and Jan vertelde het geheim aan Els (door).

The meaning of the particle can also be more or less lexicalized, as in (338a&b). Example (338b) is particularly revealing in this respect, as the particle verb overhalento persuade takes a PP headed by tot (here in the form of the anticipatory pronominal PP er toe), which can be selected neither by the verb nor by the particle.

Finally, the examples in (339) show that particles also occur in completely idiomatic constructions.

All of this suggests that in many cases the best place to account for the meaning contribution of particles is the lexicon. So, it seems best to describe the meaning of particle verbs not in a grammar but in a dictionary. We will therefore not attempt to provide further details here.