The first exceptional paradigm occurs only with (neuter) het-nouns; the deviation consists in the absence of the -e ending in definite singular noun phrases. This paradigm is illustrated in Table 3 with the collocation stoffelijk overschot “mortal remains/corpse’; the deviant case is framed in bold. We can describe this paradigm by saying that rule (3b) (i.e. [-indefinite] ⇒ adjective + -e) does not apply.

To get an impression of the robustness of the deviation from the pattern in Table 3, we performed a Google search (11/13/2022) on the two competing strings het stoffelijk overschot and het stoffelijke overschot and found that the first string is indeed more frequent than the second; cf. (20a). We also found that plural (de) stoffelijk overschotten is more frequent than (de) stoffelijke overschotten; the frequencies in the (c)-examples are based on the result after eliminating the string [de stoffelijk(e)].

This suggests that many speakers no longer interpret at least some of the relevant A+N combinations as attributive constructions, but as A+N compounds. This stands to reason, since the relevant A+N combinations form a semantic unit; they often have a specialized meaning that can be expressed by a single English word. Many linguistic terms, a small sample of which is given in (21), belong to this type.

Other cases are given in (22). We will see that for all of them one will be able to find alternating forms of the kind in (20) on the internet.

Note that this type of construction is used productively in creating names for newspapers and institutions, as shown in (23).

The rest of this subsection will show that the claim that the A+N combinations under discussion are idiomatic semantic units that are not compositionally determined is supported by several facts.

The second deviant paradigm is characterized by the absence of the -e ending in all relevant singular environments. This type of construction is only possible with nouns denoting human beings; many cases involve the names of titles or functions. The exceptional paradigm is shown in Table 4. We can describe this exceptional paradigm by saying that the rules in (3a-c) do not apply.

Examples with neuter nouns are difficult to find, because most [+human] nouns are either masculine or feminine. The easiest way to show that neuter nouns behave similarly is to add the diminutive suffix -tje to the [+human] noun, as this results in a neuter noun (sometimes with a negative connotation). Examples are given in (32) for the diminutive of werkerworker.

The relevant A+N combinations are fixed collocations. That their meaning is not compositionally determined can be supported by the fact, illustrated in (33), that the adjectives do not allow modification and cannot be used in predicative position. However, this may also be due to the fact that the adjectives in question are mostly relational adjectives, which are characterized by these properties anyway; cf. Section 23.3.3. More reliable evidence for the claim that the A+N combinations are idiomatic in nature will be provided in the following subsections.

Before doing so, however, we want to note that the uninflected plural forms in the paradigm of Table 5 are in competition with their corresponding inflected forms; cf. Haeseryn et al. (1997: §6.4.1.3). This is illustrated by the results of a Google search (March 2020) in (34); the numbers indicate the number of uninflected/inflected forms. In fact, the same may be true for the singular nouns in the primeless examples, although we will see that the alternation between the uninflected and the inflected adjective may go hand in hand with a change in meaning.

The third and final deviant paradigm is also limited to [+human] nouns and occurs mainly with nouns denoting professions of a certain social status. The deviation consists in the absence of the -e ending in the indefinite singular. Perhaps this paradigm occurs with both de and het-nouns, but since the -e ending does not occur in indefinite singular neuter noun phrases anyway, this cannot be determined. The paradigm is presented in Table 6; the exceptional case is framed in bold lines.

We can describe the paradigm in Table 6 by saying that rule (3a) (i.e. [-neuter] ⇒ adjective + -e) does not apply. Some more examples of this kind are given in (39); the primeless examples are the special form.

In contrast to the previous cases, the meaning of the noun phrase is compositionally determined; the adjective and the noun do not form a fixed semantic unit. That the adjective really denotes a property of the head noun is clear from the fact that the adjective can be modified by a degree modifier or appear in its comparative form, as the (a) and (b)-examples in (40). The superlative form is also possible, but the (c)-examples show that the noun phrase then has a definite determiner and the -e ending is present.

The constructions under discussion can usually only refer to female persons if the head noun is morphologically marked as feminine by an affix. This is demonstrated in (41): the simple noun vrouwwoman leads to a marked result in this construction, while nouns derived with the feminine affixes -e and -ster are fully acceptable.

Two notes are in order here. First, the noun taalkundige may be an exception to the general rule that only morphologically derived feminine nouns can enter the construction, although one should consider the possibility that this noun can have the feminine –e ending (cf. Section N14.3.3, sub II), but that it is not phonetically realized because the stem already ends in a schwa. Second, it is certainly not the case that all nouns denoting male persons can be used in this construction. This can be illustrated by the examples in (42), which show that the restrictions are often rather idiosyncratic.

The semantics of the examples in (40) and (41) is peculiar in that a noun phrase such as een knap taalkundigea clever linguist does not refer to the intersection of the sets denoted by the noun taalkundige and the adjective knap; cf. the discussion in Section 23.3.2.1, sub I. Instead, the adjective provides an evaluation of some property or skill that is typical of the entity denoted by the noun: een knap taalkundige does not denote a linguist who is clever in general, but a linguist who is clever as a linguist. This is also reflected in the entailment relations shown in (43); cf. Alexiadou et al. (2007). In (43a) the predicatively used noun phrase een grote jongen has an intersective interpretation, and from this we can conclude that the property denoted by the adjective also applies to the subject of the copular construction. In (43b), on the other hand, the predicatively used noun phrase een goed spreker has a non-intersective interpretation, and the entailment clearly does not hold.

In many cases, the non-intersective meaning can also be expressed by the inflected adjective, which means that the primeless examples in (44) are in fact ambiguous: for example, een vlotte typist as (44c) can refer to a typist who is sporty, or to a typist who is skilled as a typist, while een vlot typist (44c') has only the latter reading.

If more than one adjective is present, the concord constraint on attributive inflection in (5) must be respected; either the adjectives are all inflected or they are all uninflected.