Section 24.4 briefly discussed the pronominalization of the adjective (phrase). Some of the examples given there are repeated here as (322). In (322a) the pronoun hetit performs the same function as the full AP bang voor hondenafraid of dogs, while in (322b) it replaces the smaller phrase bangafraid, which is evident from the fact that the PP voor spinnenof spiders functions as the PP-complement of the pronominalized adjective.

Pronominalization of adjective phrases is also possible when a degree modifier is present. We will see, however, that many aspects of this phenomenon require further investigation in order to arrive at a clearer and more coherent description of the relevant facts.

The examples in (323) are completely parallel to those in (322); in (323a) the pronoun het performs the same function as the full AP vrij bang voor hondenrather afraid of dogs, while in (323b) it replaces the smaller phrase vrij bang. In both cases the result is perfectly acceptable.

The pronominalizations in (323) do not seem to exhaust the possibilities; two alternatives are given in (324a&b), which may be marked, but do not seem unacceptable. The pronoun het in (324a) seems to replace the smaller phrase bang voor honden, since the degree modifier zeervery is interpreted as a modifier of this phrase. The pronoun in (324b) only replaces the adjective bang; its PP-complement and its degree modifier are both overtly realized.

Not all modifiers can occur in the pronominalization construction; use of the degree modifier heelvery in (325a) gives an unacceptable result. However, the construction becomes fully acceptable if we add the adjective erg, as in (325b).

At first glance, example (325b) appears to be a special case of (324a). Since the degree modifier erg can usually be modified by the degree modifier heel, as in heel erg bang, (325b) might simply contain the complex degree modifier phrase heel erg. However, there are reasons to believe that this is not the correct analysis, and that erg in this example does not function as a degree modifier, but as a dummy element (comparable to English much), which licenses the presence of the degree modifier heel. This will become clear when we consider the examples in (326a&b) with the modifiers tetoo and zoso (cf. Section 25.1.3); these cases show that it is usually impossible to replace the AP bang voor honden with the pronoun het (except that Hij is het eigenlijk te with heavy accent on te is impeccable). However, as in (325), the structure can be saved by inserting ergmuch after the modifier tetoo/zoso.

In the cases in (326a&b), erg cannot be analyzed as a degree modifier for the simple reason that it cannot easily co-occur with te and zo in (327), although such cases occasionally occur on the internet (especially with zo). Since the use of te/zo erg in (326a&b) leads to perfectly acceptable results, erg seems to serve some as yet unknown function in these examples.

A more or less identical conclusion can be drawn from pronominalization constructions with the interrogative modifier hoehow; erg must be added to the second conjunct in (328b), despite the fact that (328a) shows that the string hoe erg bang is marked compared to the fully acceptable string hoe bang (although such strings do occur on the internet and are given as acceptable in Corver 2018). Note in passing that the exclamative hoe does not require the addition of erg: cf. Jan is bang voor honden. En hoe!Jan is afraid of dogs. And how!.

The fact that the use of hoe erg in example (328b) is faultless again suggests that erg does not act as a degree modifier; rather, it must have some other, as yet unknown function in the context of AP-pronominalization. Finally, note that not all modifiers require the addition of erg in the pronominalization construction; for instance, the cases in (324) are relatively good without erg; the examples in (329) show that the addition of erg actually has a degrading effect.

The cases in (330) show that the addition of erg is also prohibited in the case of the (complex) modifiers genoegenough and voldoendesufficiently.

The discussion above has made it clear that in some cases the addition of erg is necessary to allow pronominalization of the adjective, while in other cases it leads to (more) degraded results. At this point we want to give a speculative account of this fact, based on the analysis of comparable English cases with much in Corver (1997a/1997b). The analysis assumes that the distribution of erg is determined by the relative position of the modifier in the structure. One of the standard assumptions in generative grammar is that modifiers occupy a position external to the immediate projection of the head of the phrase, which contains the complements of the phrase. This would mean that the base structure of a modified AP is as given in (331), where PP stands for PP-complement; cf. Section 26.3.

However, the modifiers of the adjective are also ordered among themselves, suggesting that a more articulate structure is needed to provide a full account of the internal organization of the AP. For example, there is reason to think that an example such as (332a), which involves modification by the complex modifier zo ... mogelijkas ... as possible, has the structure in (332a'); the adjective is moved from its base position to a position preceding the element mogelijk; cf. also the discussion of example (151) in Section 26.3.1. If so, the corollary would be that the modifier zo is even further to the left. Furthermore, by analogy, we can assume that example (332b) has the structure in (332b').

Now assume that the head of the projection labeled YP must somehow be filled when the AP is pronominalized. If the modifier is itself part of YP, this requirement is trivially satisfied, which may explain why the addition of erg leads to a degraded result in the examples in (333).

However, if the modifier is inserted as part of the projection labeled XP (i.e. external to YP), the element erg must be inserted as a dummy to fill the head position of YP; this explains the examples in (326a&b). It will be clear that the analysis sketched here, as well as the classification of modifiers that is implied by it, will require further testing in the future.