Chapter 32 has shown that locational PPs typically occur in the complement of the verbs of location in (38a) and the verbs of change of location in (38b). We can therefore test whether locational PPs can be replaced by a pronominal PP by placing the pronominal form in the frames NP Vloc PP and NP Vloc NP PP.

Directional PPs typically occur in the complement of verbs of traversing, as in (45a). These verbs differ from the corresponding activity verbs in that they do not take the perfect auxiliary hebbento have, but the perfect auxiliary zijnto be. This allows us to test whether a directional PP can be replaced by a pronominal PP by placing the pronominal form in the PP-position in frame (45b).

It seems that most of the directional PPs from Table 17 in Section 32.3.1.2, sub IV, cannot be part of a pronominal PP. This is illustrated for the directional preposition naarto in the (a)-examples in (46); the (b)-examples show that a pronominal PP based on the directional circumposition naar ... toe is much preferred.

A similar contrast, with perhaps slightly less sharp judgments, can be observed in (47) with the directional preposition overacross and the directional circumposition over ...heenacross.

Note in passing, that the directional prepositions naar and over differ sharply in this respect from the functional prepositions naar and over in (48), which can easily occur as part of pronominal PPs.

Examples (49a-b) seem to show that, like naar and over, the directional prepositions vanfrom and viavia do not allow the formation of a pronominal PP. Since there is no circumpositional variant available in this case, and since (49c) shows that the form with a [-R] pronoun is not possible either, there seems to be no pronominalized counterpart for (49a) at all. Note, however, that because replacing Utrecht with one of the pronouns in (49c) is never possible, there is no reason to think that this example is ungrammatical: if, for example, hem refers to a person, as in Marie is van/via Jan/hem (naar Groningen) gereden, the result is perfectly acceptable; R-pronominalization of van/via Jan is not possible either, but this can of course be attributed to the [-human] constraint discussed in Section 36.1.

For completeness’ sake, it should be noted that the intended propositions can be expressed by using the deictic locational forms hierhere and daar ‘there to replace Utrecht in (49a); note that these forms are not [+R] pronouns, as is clear from the fact illustrated in (50) that they cannot precede the preposition.

The fact that R-pronominalization of PPs with a proper noun (here: Utrecht and Jan) is excluded for independent reasons does not mean that we have to leave the issue regarding the acceptability of R-pronominalization undecided: the examples in (51), in which the nominal complement of vanfrom and viavia is headed by the lexical noun stadcity, leave no doubt that R-pronominalization is impossible.

At first sight, the directional preposition voorbijpast in (52) seems to occur in a pronominal PP, but we probably have to put this example aside as irrelevant; the fact that hierhere can be omitted indicates that it does not function as the complement of the preposition voorbij but as an independent adverbial phrase, so that voorbij in this example is actually an intransitive preposition or a verbal particle; in short, hier cannot be analyzed as a [+R] pronoun.

The prepositions langsalong and doorthrough, which can be used both as locational PPs and as directional PPs, seem to allow R-pronominalization not only in the locational constructions in (39c&d), but also in the directional constructions in (53). Note that the use of the pronominal PP er door in (53b') has more or less the same degree of acceptability as the use of the pronominal circumpositional phrase er door heen.

The examples in this subsection have shown that directional prepositions differ in whether or not they can occur in pronominal PPs. Some reject this categorically, some seem to be ousted by their corresponding circumpositional forms, and some allow it without any problem.

Section 32.3.3, sub II, looked at several prepositional phrases which are not selected by the verb (or the head of some other phrase of which they are part), but introduce a referent that plays a certain semantic role in the clause (or other relevant phrase). Such PPs generally allow the formation of a pronominal PP; this is illustrated below for the individual prepositions that can head such phrases.

PP-complements of verbs, adjectives, nouns and adpositions all allow R-pronominalization. Since we will discuss this in more detail in Section 36.3.2, sub I, we will give only one example of each case here.

Adverbial phrases headed by a non-spatial/temporal preposition cannot undergo R-pronominalization, i.e. the prepositions in Table 30 in Section 32.3.3, sub IIC cannot head a pronominal PP. The last column of Table (67) shows that these prepositions cannot be followed by a weak pronoun either, regardless of whether the pronoun is [-human] or [+human]. The examples with dankzij, namens, ondanks, vanwege, volgens and zonder become acceptable if the weak pronoun is replaced by a phonetically strong [-neuter] one, i.e. the pronoun hemhim, haarher or henthem, which are always used to denote a [+human] entity.

It is not immediately clear how the unacceptability of pronominal PPs in this table should be explained.. One option would be to appeal to the adverbial status of the PPs in question. Another alternative approach is to consider that, at least diachronically, many of the prepositions in question are morphologically complex (or non-native); cf. Zwarts (1997a) and Corver (2022b:44) for discussion and more (formal) examples. Following up on this proposal, we should note that there are possible counterexamples to the generalization that non-spatial/temporal PPs do not allow R-pronominalization; cf. daaromtrentthereabout and hieromtrenthereabout. However, since these formations are only used in formal language, it might be better to consider them fossilized lexical items. For completeness’ sake, note that the formal forms daarentegenon the other hand and daarenbovenmoreover do not seem to be relevant here, as they are based on the medieval (simple) prepositions entegen and enboven; cf. etymologiebank.nl. We will leave this issue for future research. Note that the restriction on R-pronominalization that the preposition is morphologically simplex does not imply that it is always monosyllabic: cf. erboven ‘above it’ and eronder ‘under it’.