Generic interpretations of definite noun phrases are not encoded in any part of the noun phrase itself, but depend on the semantic content of the construction in which they occur. An example such as (46a) does not trigger a generic interpretation, since it is highly unlikely that a stage-level property such as being in a cage is a property of (the prototype of) the set of entities denoted by a noun such as zebrazebra. Therefore, this sentence must be interpreted as a proposition about a specific entity in domain D. On the other hand, an example such as (46a'), which involves the individual-level predicate of being striped, can be seen as a general statement about (the prototype of) this set of entities. The noun phrase de zebra can therefore be given both a generic and a referential interpretation. Note, however, that an example such as (46a') is ambiguous only on paper. Leaving aside contrastive accent, the two interpretations are distinguished by accent: in the referential reading of the noun phrase, the main accent is given to the adjective gestreept; in the generic reading, the main accent falls on the noun phrase (i.e. the noun zebra). A similar difference can be seen in (46b&b').

The above discussion does not imply that the generic interpretation of definite noun phrases is completely determined by context. This becomes clear when we consider some more examples. All the primeless examples in (47) would be conceivable as generic statements, which is clear from the fact that the primed examples involving indefinite noun phrases actually have the intended meanings. Nevertheless, they strongly favor a regular referential meaning, i.e. they are preferably construed as assertions about a certain entity in domain D.

The reason for the impossibility of the intended generic readings of the primeless examples is not entirely clear. It may be that we are simply not inclined to imagine a prototypical member of the sets denoted by the nouns in (47). While the noun vrouwwoman or zebra easily evokes a prototype, nouns like meisjegirl, boekbook or braadpanfrying pan do not. Perhaps this suggestion can be supported by the fact that a prototypical reading can be evoked if the context provides sufficient clues that such a reading is intended. This is clear from the fact that the examples in (48) allow for a generic reading because the syntactic context makes it likely that two prototypes are being compared: in (48a) a prototypical girl of a certain age and a prototypical boy of the same age, and in (48b) a prototypical girl from the polder and a prototypical girl from the city. However, it seems that even in these cases the use of an indefinite noun phrase, as in the primed examples, is much preferred by most speakers.

The generic interpretation of the noun phrases in the primeless examples in (48) is clearly facilitated by the use of the modifiers: in (48a) by the use of the adverbial phrase op die leeftijdat that age and in (48b) by the attributively used PP uit de polder/stadfrom the polder/city. That attributive modifiers can facilitate the generic reading is also clear from the examples in (49). Perhaps the use of the attributive modifier gebonden facilitates a prototypical reading because it divides the superset of books into two subsets, so that we can compare the prototypical members of these subsets: the prototypical member of the set of bound books is unaffordable, in contrast to the prototypical member of the set of paperbacks or pocketbooks.

This probably also explains why classes that are relatively high in the speaker’s taxonomy are usually not preceded by a definite article in generic sentences. Acceptability judgments on examples such as (50a) contrast sharply with those on examples such as (46b): the fact that mammals are higher in the taxonomy than zebras apparently makes it easier for the speaker to imagine a prototypical zebra than a prototypical mammal. Referring to a typical member is easier, so the use of an indefinite article, as in (50b&c), is preferred.

From the discussion above we may perhaps conclude that the ambiguity between the regular referential reading and the generic reading of a singular definite noun phrase is related to the question whether the language user is able to interpret the noun phrase as referring to a prototype of a certain set of entities (where many non-linguistic aspects may play a role).

Let us now consider whether plural definite noun phrases are also possible in generic statements. The (a)-examples in (51) can only be interpreted as statements about a contextually determined group of zebras/women. An example such as (51b) seems to fare better as a generic statement, but this is due to the fact that the NP grote katbig cat can be used as the name of the superset that contains the subsets of cats denoted by the nouns leeuwlion, tijgertiger, etc. In other words, the noun phrase de grote katten does not refer to one but to several species of animals, hence its plural form.

This seems to lead to the conclusion that plural definite noun phrases cannot be used as generic noun phrases unless the noun phrase denotes a set of entities that can be further divided into several conventionally distinguished subclasses/species. This conclusion seems more or less correct, but it turns out that we need to make at least one exception. Consider example (52), taken from Geerts et al. (1984), which contains the same string of words as (51a), but which seems to be perfectly fine on a generic reading. The crucial ingredient of (52) that makes the definite determiner felicitous is the presence of the focus particle alleenonly; as soon as it is deleted, the output becomes bad on a generic reading. The focus particle alleen seems to be able to license the use of the definite article due to the fact that it evokes a reading in which the set denoted by zebra is construed as a proper subset of a larger set, viz. the set denoted by wilde paardenwild horses.

A similar effect on the legitimacy of a definite determiner in generic plural noun phrases can be seen in the case of restrictive attributive modifiers. This can be illustrated by the examples in (53), adapted from De Hoop, Vanden Wyngaerd & Zwart (1990:100ff.); the semantic effect of the addition of the PP-modifier met witte voetjes is to create a subset of domestic cats with specific physical characteristics (viz. the possession of white paws), and as a result the definite determiner can now be felicitously used to pick out the intended subset.

The minimal pair in (54) shows that it is only the subset that can occur with the definite article; in (54a) the noun phrase de katten refers to a superset that includes the subset referred to by de katten met witte voetjes, and the use of the definite article leads to a degraded result, whereas in (54b) the noun phrase de zwarte katten refers to a subset that is contrasted with another subset referred to by de katten met witte voetjes, and the use of the definite article is allowed.

From this we can conclude that the use of the definite article is not related to the fact that the noun phrases in question have a generic reading, but to the fact that these noun phrases are linked to some explicitly mentioned or tacitly assumed superset in domain D. This use of the definite article is therefore reminiscent of the use of the definite article in noun phrases that refer to entities that are not part of domain D but can be inferred from the linguistic or non-linguistic context of the discourse; cf. the discussion of examples in Section 18.1.1.2, sub IIIC.

Although the discussion in Subsection 2 has shown that plural definite noun phrases cannot normally be used generically, an exception must be made for nationality nouns like NederlanderDutchman or nouns that refer to members of certain societal groups or organizations like kapitalistcapitalist. For these types of nouns, generic statements can therefore often be expressed in four different ways, as shown in (55) and (56).

So far we have only discussed generic definite noun phrases in clauses. As shown in (57a), definite noun phrases can also get a generic reading when they are embedded in a larger noun phrase. The difference between (57a) and (57b) suggests that here too the context determines whether a generic reading is possible or not.

Note that (57a) is only genuinely ambiguous on paper; when pronounced in a neutral context, the generic reading leads to main stress on vrouw, while the specific reading assigns main prosodic prominence to rechten. This is shown in (58), where the verbal predicate blocks a generic reading of de vrouw in (58a), but strongly favors it in (58b).

The contrast between the examples in (57a) and (59) shows again that definite plural noun phrases are not normally assigned a generic reading.

Generic indefinite noun phrases differ from non-generic ones in their syntactic behavior. Consider the examples in (61). Non-generic indefinite DPs headed by an indefinite article do not occur in the regular subject position, whereas generic indefinite noun phrases introduced by the article een/∅ must occur in this position, which is clear from the fact that they cannot enter the expletive construction discussed in Section 21.1.2; the noun phrases in (61a&a') receive a non-generic interpretation, whereas those in (61b&b') receive a generic interpretation.

It may be, however, that an exception must be made for generic statements of the type in (62). These examples are generic, but not in the same sense as the examples discussed earlier: they do not involve a categorical statement about the members of the set denoted by the NP goed mesgood knife, but a generic statement about the activity denoted by the noun phrase dit soort werkthis kind of work; if one does this (kind of) work, a good knife is/good knives are indispensable. Therefore, if we want to categorize the subject noun phrases in (62) as non-generic, we should rephrase our earlier findings a bit: it is only in generic clauses that indefinite noun phrases introduced by een/∅ can occupy the regular subject position. Since, to our knowledge, examples such as (62) have not been discussed in the literature, we will not consider them further.

So far we have not discussed the difference between singular and plural generic indefinite noun phrases. Although it may seem difficult to pinpoint a difference in meaning, it is clear that they are not synonymous. This becomes clear when we consider the implication relations that hold between singular and plural examples, as in the primeless and primed examples in (63).

It seems that implication relations like (63a) are always valid. The inverse implication relation in (63b), on the other hand, does not seem to hold in all cases. This suggests that generic sentences with an indefinite singular noun phrase express that the typical member of the class is in some sense inherently endowed with or defined by the property denoted by the predicate. Generic sentences with an indefinite plural noun phrase, on the other hand, seem to ascribe a more accidental or transitory property to the class: for example, musicals may be popular today, but there is no guarantee that this will also be the case in the future. That something like this is indeed the case is shown by the fact that the use of an adverbial phrase such as tegenwoordignowadays is possible in (64a), but not in (64b).

There are also differences in the syntactic environments in which singular and plural generic indefinite noun phrases can occur. Above we have only discussed examples where the generic noun phrase acts as the subject of a clause. If we extend our discussion to other syntactic functions, it seems that generic singular noun phrases have a more limited distribution than generic plural ones. The primeless singular examples in (65) must be interpreted as specific: Jan studies/is fond of a certain zebra. The singly-primed plural examples, on the other hand, seem to easily allow a generic interpretation of the indefinite noun phrase. The doubly-primed examples are added to show that generic definite noun phrases can also be used in syntactic functions other than subject.

Note, however, that we cannot conclude from these examples that generic singular indefinite noun phrases can only occur as the subject of the clause. This is clear from the examples in (66). In these examples the indefinite noun phrases are not the subject of the clause, but the sentences can still be interpreted generically (which is made easier if a modifier like meestal is added to the sentence). The main difference between the examples in (65) and (66) is that the latter examples contain a complementive predicated of the indefinite noun phrase. From this we can conclude that a generic singular indefinite noun phrase can only occur if it is the logical subject of some predicate, whereas generic plural indefinite noun phrases are freer in distribution.

Finally, we give the primeless examples in (67) to show that generic indefinite noun phrases can also be embedded in a larger noun phrase. The meaning of these examples is something like “all zebras have a biotope that consists of ...”. As shown in the primed examples, the larger noun phrase containing a singular indefinite noun phrase also has a more restricted distribution than the one containing a plural indefinite noun phrase; (67a') is at least preferably interpreted as involving the biotope of a particular zebra.

When a definite article is combined with a substance noun like wijnwine or fruitfruit, the resulting noun phrase is often given a specific interpretation: the definite noun phrase refers to a contextually determined quantity of the substance in question. However, it is not impossible to find substance nouns with a definite determiner that receive a generic interpretation; example (68) gives some examples of both uses.

The singular indefinite article een cannot normally be combined with non-count nouns. However, the examples in (69) show that the indefinite null article ∅ can be used. If the resulting noun phrase functions as subject, its interpretation depends on its position in the clause: if the noun phrase occupies the regular subject position, as in (69a), it must be interpreted generically. If the noun phrase is in the expletive construction, as in (69a'), it is always interpreted as a non-generic, indefinite noun phrase. In other functions, the interpretation of the noun phrase depends on the denotation of the verb phrase, as can be seen by comparing the two (b)-examples.

When the indefinite noun phrase is embedded in a larger noun phrase, it is also the context that determines whether a generic reading is possible. The (a)-examples in (70) show this for substance nouns embedded in a subject, and the (b)-examples for substance nouns embedded in a direct object.

Non-deverbal abstract non-count nouns can also be construed with the definite articles de and het without necessarily receiving a specific interpretation. Again, the context determines the distribution of specific and generic readings.

Abstract non-count nouns in argument positions cannot normally be combined with the indefinite article een without triggering a special, exclamative interpretation; cf. Section 18.1.4.318.1.4.2. However, the addition of a restrictive modifier may license it: Er heerst daar een *(ondraaglijke) armoedeThere is an unbearable poverty there; Hij heeft een *(onzegbaar) verdrietHe has an unspeakable sorrow. However, as far as we can tell, such indefinite noun phrases are not easily possible in generic contexts; the examples in (72) are somewhat odd.

Abstract non-count nouns with the indefinite null article ∅, on the other hand, are possible in generic contexts. As with substance nouns, this is the normal interpretation when an indefinite noun phrase occupies the regular subject position; to obtain a non-generic reading, the indefinite subject must occur in the expletive construction.

If the indefinite noun phrase has another syntactic function in the sentence, the context determines whether a generic interpretation is possible or not; cf. the (a)-examples in (74). If the indefinite noun phrase is embedded in a larger noun phrase, as in (74b), the generic reading of the indefinite noun phrase is the most prominent one.

Here we will confine our discussion to nominal infinitives; cf. Sections 14.3.1.2 and 15.2.3.2. We start with bare-inf nominalizations (i.e. without an article). Since (75a) is derived from the intransitive, habitual verb rokento smoke, it is not surprising that such nominalizations are generally generic; cf. Jan rooktJan smokes, i.e. Jan is a smoker. The same is true for (75b), since the nominalized phrase contains a bare plural noun, which in this example seems to be construed generically; cf. Jan rookt sigarenJan smokes cigars, i.e. Jan is a smoker of cigars.

Det-inf nominalizations (preceded by the neuter article het) can also inherit the arguments of the verb. The examples in (76) show that when the arguments precede the infinitive, they must be realized as indefinite plural noun phrases, just as in the case of bare-inf nominalizations, but in this case the nominalizations clearly refer to specific “smoking” events, so we can safely claim that we are dealing with non-generic uses of these noun phrases.

The examples in (77) further show that when we place these noun phrases in contexts that favor a generic interpretation, the result is marginal at best.

If the argument follows the infinitive, there are no restrictions on its realization; it can be plural or singular, and it can be indefinite or definite. The (a)-examples in (78) show that the choice between the latter two options affects the interpretation of the noun phrase as a whole: if the argument is a bare plural noun phrase, the generic reading of the nominalized phrase is clearly favored; on the other hand, if the argument is definite, a generic reading seems to be blocked. The (b)-examples show that, as a result, the use of a definite argument is excluded in contexts favoring a generic interpretation.

Note, however, that definite noun phrases introduced by a demonstrative are possible in contexts like (78b&b'): if the noun phrase is singular, a specific reading is triggered; if it is plural, both readings are available. For completeness’ sake, example (79c) shows that substance nouns such as tabaktobacco also allow both readings when preceded by a demonstrative.

Finally, consider cases where nominalized phrases are embedded in a larger noun phrase. Given the contrast between the examples in (75) and (77), the judgment on example (80a) is surprising. This example clearly has a generic interpretation, but nevertheless the nominalized phrase must be preceded by the definite article. The judgments on the remaining examples (80b&c) are consistent with the judgments on the examples in (78).