That we are dealing with three different but homophonous noun types in (127) is clear not only from the meaning differences between the three constructions, but also from the fact that the nouns have different genders. Consider again the examples in (127a&b): the noun soortspecies in (127a) is non-neuter, which is clear from the fact that it takes the non-neuter demonstratives deze/diethis/that: deze/die soort hond/honden. The noun soortkind of in (127b), on the other hand, is neuter, which is clear from the fact that it takes the neuter demonstratives dit/datthis/that: dit/dat soort auto/auto’s. It is difficult to determine the gender of the noun soortN-like entity in (127c), since it differs from the other two nouns in not allowing these definite demonstratives at all.

The non-neuter noun soortspecies imposes gender restrictions on N2; it can be followed by singular non-neuter nouns such as honddog, but not by singular neuter nouns such as paardhorse. The neuter noun soortkind of does not impose similar restrictions on N2, which can therefore be either neuter or non-neuter. The third use of soort is also compatible with both neuter and non-neuter N2s. The pattern given in (129) is supported by a Google search (March 5, 2020): all strings in (129) occur at least several times in the intended construction, except for deze/die soort paard (only a single attestation with deze).

Non-neuter N1s like kleurcolor and maatsize in the primeless examples in (130) often behave like the non-neuter noun soortspecies in that they do not allow neuter N2s. Neuter nouns like formaatsize and genretype in the primed examples do not impose a similar restriction on N2s; cf. Schermer-Vermeer (2008). However, it seems that the status of mixed cases with non-neuter N1s and neuter N2s also depends on the choice of N2: an example such as deze/die kleur hemd in (130a) is less common than deze/die kleur behangthis/that color of wallpaper in (128a) or deze/die kleur haarthis/that color of hair, both of which involve a neuter N2 but can be easily found on the internet. Perhaps the distinction between count and non-count nouns also plays a role; we leave this to future research.

Another striking difference between the examples in (127) is that the non-neuter noun soortspecies in (127a) can occur as the second member of a compound with a similar kind of meaning, whereas this is completely impossible with the N1 soortN-like entity in (127c); the compound appelsoort in (131c) is of course acceptable but only as the counterpart of the binominal construction dat soort appelthat species of apple. Example (131b) further shows that the result with the neuter N1 soortkind of in (127b) is marginal (although phrases such as het hondensoort can be found on the internet); moreover, it seems difficult to interpret this compound with the “resemblance” reading typical of this noun.

The examples in (132) show that nouns like kleurcolor, kwaliteitquality and maatsize behave like the non-neuter noun soortspecies. All these examples are common on the internet.

The noun soortN-like entity also differs from the other two nouns in that it does not allow plural formation: example (133c) is acceptable, but only under the same kind of reading as (133a) or (133b); it does not have the interpretation “two apple-like things”. The plural examples in (133a) and (133b) impose different selection restrictions on N2: N2 must be plural in the former, but can be singular in the latter.

A caveat is in order here, however. For convenience, we have translated the non-neuter noun soortspecies in (127a) with the English noun species. This translation may actually be too narrow, since it can also be combined with N2s like postzegelsstamps and substance nouns like koffiecoffee.

This may raise the question as to whether we are really dealing with the plural form of the neuter noun soortkind of in (133b); it may actually involve the plural form of the non-neuter noun. One reason for assuming this is that the neuter noun cannot be modified by a quantifier like elkeach. Since nouns that have a plural form generally do allow modification by elkeach, the unacceptability of (135b) casts some doubt on the assumption that we are dealing with the plural form of the neuter noun soort in (133b). We leave this to future research.

It is difficult to determine whether nouns like kleurcolor, kwaliteitquality and maatsize behave like the non-neuter noun soortspecies with respect to pluralization; the examples in (136) show that the plural form kleuren can easily be followed by a substance noun, but not by a count noun. We leave it to future research to determine the precise status of examples such as (136).

The constructions in (127) are similar in that they normally do not allow a definite determiner. In the case of (127a&b), however, this restriction is relaxed when the construction is modified by a relative clause. The binominal construction in (137c) is acceptable, but only on a reading comparable to that found in (137a) or (137b).

Example (138) shows that a noun such as kleurcolor behaves in this respect like the non-neuter noun soortspecies.

Attributive adjectives can license the indefinite determiner een on the nouns soortspecies and soortkind but only if they precede N1. This is shown in (139); the primed examples are only acceptable under the “of a sort” reading (i.e. (139a') can be interpreted as “a beautiful dog of a sort”), in which case an indefinite article must be present..

The examples in (140) show again that a noun like kleur behaves in this respect like the non-neuter noun soortspecies

Another difference involves the insertion of the preposition vanof between N1 and N2. The examples in (141) show that this is easily possible in examples such as (127c). Examples such as (141a) sound somewhat marginal. They can be found on the internet, but the number of cases is relatively small: a Google search (March 5, 2020) for the string [deze soort van] yielded about 140 hits, but many of these did not instantiate the relevant construction. Judgments about examples such as (141b) vary between speakers, but examples of this construction are common in informal spoken Dutch and are easy to find on the internet; a second Google search for the string [dit soort van] yielded 180 hits, and a cursory look at the results revealed that most cases did indeed instantiate the relevant construction.

The examples in (142) with the indefinite article een are all acceptable, but only on a reading similar to (141c). Note that the addition of an attributive adjective triggers a more referential reading of the noun soort, which makes the example unacceptable.

The examples in (143) show that with nouns like kleurcolor, kwaliteitquality and maatsize, the insertion of van leads to an unacceptable result. These nouns therefore seem to pattern again with the non-neuter noun soortspecies.

This section has discussed some of the properties of the non-quantificational constructions in (127); it has been shown that the three homophonous forms of N2 (i.e. soort) differ in various ways. Since these forms have not been systematically studied in the literature, future research on the three constructions in question may reveal more systematic differences. This section has also shown that the behavior of the other nouns that can enter the non-quantificational construction is similar to that of the non-neuter noun soortspecies in (127a).