Unlike regular (non-modified) partitive genitive adjectives, anders can only be preceded by quantificational pronouns like iets/watsomething and nietsnothing: quantifier noun phrases and quantifiers are excluded, and the wat voor construction does not occur either. This is shown in (125).

Note that the more or less fixed expression in (126a) with the negative pronoun nietsnothing is somewhat special in this respect in that the quantifiers veelmuch and weiniglittle can also occur. Example (126b) shows that the quantifier veel must be preceded by the negative marker nietnot. The quantifier weinig in (126c) is inherently negative, meaning “not much”; it cannot be preceded by niet, because this would cancel the inherent negation.

The element anders also differs from partitive genitives in that it can be combined with the [+animate] quantificational pronouns iemandsomeone and niemandno one; cf. Section 29.2.3, sub I. In fact, this does not exhaust the possibilities, as it can also be combined with the place adverbs ergenssomewhere and nergensnowhere, neither of which occur in the partitive genitive construction.

If the [-animate] pronoun (n)iets occurs as the complement of a preposition, R-pronominalization is possible and sometimes even preferred. The two alternative realizations are given in (128).

The examples in (129) show that the phrase (n)iets anders behaves completely on a par with the quantificational pronoun.

The acceptability of (129b) is quite remarkable in view of the fact that R-pronominalization is usually impossible when the pronoun following the preposition is part of a larger phrase. This is illustrated in (130) for cases in which the preposition is followed by a partitive genitive construction.

The contrast between (129b) and (130b) again suggests that the phrase (n)iets anders is not a partitive genitive construction (regardless of an explanation for the remarkable acceptability of (129b)).

The construction iets anders can be modified by the degree modifier heelcompletely. The primeless examples in (131) show that this modifier can be placed either after or before the quantificational pronoun iets without a clear difference in meaning. The topicalization construction in (131a'), which requires contrastive accent because the phrase iets anders is indefinite, shows that the string heel iets anders behaves as a single constituent; cf. the constituency test. Note that heel can be replaced by the near-synonymous adjective totaalcompletely, but we will not illustrate this here.

The pronoun iets usually alternates with wat, and at first glance the examples in (132) suggest that this is also possible here, but we will see that there are at least two small differences between the two sets of examples in (131) and (132).

First, the (a)-examples with prepronominal heel differ in meaning. Although (131a) and (132a) can both be understood with the modifier heel as a degree modifier of the adjective anders as “something quite different”, example (132a) with wat allows an additional reading in which heel is a modifier of the quantificational pronoun “quite a lot of different things”. The fact that heel cannot be construed with iets in (131a) is of course related to the fact that we see the same difference when iets and wat are used as independent arguments: heel wat/*ietsquite a lot.

Second, example (131b) with iets is perfectly acceptable, while the similar construction with wat in (132b) seems to yield a poor result (although such examples can be found on the internet). The contrast is perhaps even sharper when the noun phrases are used as subjects: a Google search (January 2023) on the string [er stond iets heel anders] resulted in 12 hits, while corresponding string with wat did yielded no result (apart from references to the present work).

The conclusion that prepronominal heel can be construed with wat but not with iets also explains the contrast in (134). Example (134a), in which wat is both preceded and followed by an occurrence of heel, seems marginally possible if the first occurrence of heel is construed as a modifier of the nominal part and the second one as a modifier of anders. Example (134b), on the other hand, is unacceptable because the first occurrence of heel cannot be construed with iets and must therefore be interpreted (redundantly) as a modifier of anders.

The main observation for our present discussion is that the prepronominal modifier heel is able to modify the adjectival part anders; we can now show that the partitive genitive constructions such as those in (135) behave quite differently in this respect. Note that while (135a) is fine with or without heel, (135b) does not tolerate heel. The unacceptability of (135b) with heel suggests that the partitive genitive following the pronoun cannot be modified by prepronominal heel. This is also supported by the fact that (135a) has only one reading, in which heel is construed as a quantifier of wat, resulting in the reading “quite a lot”. In order to construe heel as a degree modifier of the adjective, it must be placed to the right of the pronoun, as in (135c).

We can now safely conclude that cases like iets/wat anders are not partitive genitive constructions. Note that the examples in (136) show that in constructions with iemandsomeone and ergenssomewhere, the modifier heel may not even follow the noun, but must immediately precede it.

The examples in (137) show that the same thing holds if ergens arises as a result of preposition stranding: while heel can either precede or follow iets in (137a), it must precede ergens in (137b).