Section 15.3.1.3, sub IVA, has shown that intransitive verbs do not allow ing-nominalization: the verb huilento cry, for example, has no corresponding Ing-noun *huiling. Input verbs for ing-nominalization should have an internal argument.

Unaccusative verbs easily accept ing-nominalization. The examples in (339) show that the theme argument must normally be expressed in the form of a postnominal van-PP or a prenominal genitive noun phrase/possessive pronoun. In the latter case, the theme argument must be [+human].

The cases in (340a&b) show that leaving the theme argument unexpressed leads to questionable results even in generic contexts; apparently, it is difficult in such cases to give the unexpressed theme a non-specific interpretation. Example (340c) shows that the prenominal genitive noun phrase/possessive pronoun and the postverbal van-PP cannot co-occur because they both express the theme argument; like their unaccusative input verbs, these ing-nominalizations assign the theme role to only one argument.

There are a limited number of cases in which it looks as if the theme of the corresponding verb is realized as an attributive adjective. These cases are limited to relational adjectives of the geographical type (cf. Section A24.3.3), like AmerikaansAmerican, Amsterdamsof Amsterdam, etc.

However, the analysis proposed here is by no means uncontroversial: the fact that relational adjectives differ from other adjectives in that they do not denote a property but express a relation between two entities does not mean that they are to be interpreted as the inherited theme argument of the input verbs opkomento rise and bloeiento flourish; instead, it can be argued that they fulfill the same function as in de Amerikaanse dollarthe American dollar or de Amsterdamse grachtenthe Amsterdam canals, where they cannot be seen as arguments of the nouns.

Ing-nominalizations based on transitive verbs offer a wider range of possible forms of complementation. Two common uses can be distinguished: one in which both arguments are expressed, and one in which only the theme argument is expressed. Let us start with the latter type of construction.

This subsection considers triadic ing-nominalizations derived from ditransitive transfer verbs like uitreikento present and overdragento transfer/hand over. As with inf-nominalizations, it is in principle possible for ing-nominalizations to occur with all three arguments, although in practice this is very rare; more often one (typically the agent) or two (agent and recipient) of the arguments are left unexpressed. In general, the presence of the theme argument is required, although it can be omitted in generic statements like (351).

The following subsections describe the cases in which one or more arguments appear.

Ing-nominalizations can also inherit PP-arguments from input verbs like jagen opto hunt (for). The examples in (359) show that the ing-nominalization inherits the preposition selected by the input verb: the theme does not appear as a van-PP, but as a PP headed by op. These examples also show that it is easier to place the agentive door-phrase in front of a PP-complement than in front of a theme realized as a postnominal van-PP; cf. (346). Again, this may have to do with the fact that these PP-complements may undergo PP-over-V in the corresponding verbal construction.

Another difference between these ing-nominalizations and those derived from transitive verbs is illustrated by the examples in (360), which show that in dyadic constructions involving inheritance of a PP argument, only the agent argument can appear as a possessive pronoun/genitive noun phrase, suggesting that the selected preposition op must be overtly realized.

The inherited PP-argument need not be a theme; the nominalization of the verb aanbevelen voorto recommend for in (361), for example, involves a third argument of the verb preceded by the preposition voor (and to which, for convenience, we assign the thematic role of goal). The (a)-examples show that the theme argument prefers to precede the goal argument, and the (b)-examples illustrate again that it is easier to place an agentive door-PP like door de commissieby the committee in front of the inherited PP-complement than in front of a theme realized as a postnominal van-PP. Example (361c), which combines the two disfavored orders in (361a') and (361b''), is severely degraded.

The examples in (362) show that also in this case the agent or theme argument of the ing-nominalization can appear prenominally as a genitive noun phrase/possessive pronoun; the goal argument always appears as a postnominal PP.

Unlike inf-nominalizations, ing-nominalizations usually do not accept verbs that select an adjectival complementive as input. The primed examples in (363) show that these constructions are unacceptable, regardless of whether the predicate is postnominal or prenominal.

Note, however, that compound-like forms such as veiligverklaringsafety declaration are common in more formal texts. Such forms are perceived as compounds: most dictionaries include such common forms as heiligverklaringcanonization/beatification and goedkeuringapproval. The greater degree of acceptability of such compounds in certain formal registers (religion, health, finance, legislation, etc.) may be related to the fact that the combination verklaren + adjective (heiligsacred, dooddead, faillietbankrupt, gegrondjustified, etc.) can denote actions typically belonging to the profession. This may suggest that the adjectives involved are interpreted as a kind of verbal particle, which can also be part of compound-like ing-nominalizations found in dictionaries; cf. onderdompelingimmersion, derived from the particle verb onderdompelenimmerse.

When the complementive is introduced by a preposition such as totto or alsas, ing-nominalization is also possible, as shown in (364a&b). In such constructions the complementive can only occur postnominally; placing it in prenominal position results in unacceptability.

The previous subsections have dealt with the most important aspects of complementation of ing-nominalizations, in particular the form and position of the various arguments and their relation to the derived noun. Let us summarize the main results. In unaccusative ing-nominalizations, the theme argument is obligatory and typically appears postnominally as a van-PP. This also holds for the theme argument of dyadic ing-nominalizations (at least in non-generic contexts); agents can be expressed as door-PPs and usually follow the theme, but are optional. In triadic ing-nominalizations, themes are again typically obligatory, while recipients and agents are often omitted. If the latter are expressed, they are realized as aan and door-PPs, respectively, and follow the theme in postnominal position. The agent usually follows the recipient (if present). In all cases, the theme and agent arguments can also take the form of a prenominal genitive noun phrase or a possessive pronoun, provided that they are [+human]. All this can be represented as in Table 11, which gives us the basic patterns of ing-nominalizations; the table does not include ing-nominalizations derived from verbs taking a PP-complement or a complementive introduced by als/tot, which are also inherited by the nominalization.

Ing-nominalizations can be seen as inheriting the argument structure of the input verb, with the nominal construction resembling the verbal construction in the number of arguments and their thematic functions. However, while the arguments of verbs are usually expressed explicitly, this need not be the case for the inherited arguments of the corresponding ing-nouns. If the ing-nominalization is derived from a transitive verb, as in (365a), the theme must be present, while the agent can easily be omitted. If the input verb is ditransitive, the recipient need not be expressed either, as in (365b).

However, even if they are omitted, the agent and recipient arguments are still implied and must be recoverable or inferable from the context. Under such circumstances, it may even be possible to omit the theme, although this is more likely to lead to a marked result. As shown in (366), ing-nominalizations derived from verbs selecting a PP-complement essentially pattern with those derived from (di-)transitive verbs; the theme argument is usually realized.

The van-PP of ing-nominalizations cannot occur in post-copular position, as shown in (367a-c). This is to be expected, because van-PPs in post-copular position are usually interpreted as possessive elements, whereas states of affairs, the denotation of ing-nominalizations, cannot be possessed. For completeness, example (367d) shows that PPs introduced by a preposition other than van cannot be used in this position either.

The acceptability of the instances in (368) shows that ing-nominalizations allow R-pronominalization of theme arguments.

R-pronominalization of agents or recipients, on the other hand, is excluded; this is illustrated by the unacceptability of the examples in (369).

The PP-extraction test produces results that are far from clear. The acceptability of these sentences depends on the ease with which a contrastive interpretation can be construed. Nevertheless, there seem to be differences in acceptability that neither context nor differences in verb type or number of arguments can account for.

Table 12 summarizes the results from the previous subsections of the four tests for inherited theme arguments of ing-nouns. The third and fifth columns indicate whether the results provide evidence for or against the assumption that we are dealing with complements. The first three tests provide unequivocal evidence for complement status of both van-PPs and of PP-themes headed by other prepositions. The results of the PP-extraction tests seem to contradict this, but we have seen that these tests are problematic in various ways and may not be suitable for establishing complement status anyway. We therefore conclude that the theme functions as an argument of the derived noun.

For recipient aan-PPs and agentive door-PPs it is more difficult to establish whether they are arguments of the noun. Only the first test is relevant for these, and it seems that this test provides only weak evidence for assuming argument status: recipients and agents are semantically implied, but need not be syntactically expressed. However, since recipients and agentive door-phrases are usually optional in verbal constructions, the evidence is inconclusive. We will therefore assume that they have a status similar to that of the theme, which clearly behaves as an argument.