Consider the examples in (143), in which the direct object introduces a new entity into domain D if it is indefinite (een krant), or refers to a unique entity that is contextually given if it is definite (de krant). In both cases, the pronoun hem can take the direct object as its antecedent (which is indicated by indices).

Things are different when the definite object is adjacent to the verb in clause-final position, as in (144a). Under a neutral intonation pattern (with non-emphatic sentence stress on the object), the complex-predicate reading becomes more prominent: the use of a deictic pronoun causes the sentence to be degraded, as the sentence no longer refers to reading an actual newspaper, but only to the activity of newspaper reading. A similar case is given in (144b) with a definite noun phrase embedded in a PP-complement of the verb: this case does not normally refer to the action of looking at a specific television set, but to watching some television program(s). The number signs indicate that the weak reading of the definite noun phrase makes the referential identity indicated by the indices less felicitous.

There may be some disagreement among speakers as to whether or not the regular referential reading of the definite noun phrase de krant in (144a) is possible, but all speakers will agree that this reading is blocked in out-of-the-blue contexts: example (145a) can be uttered unexpectedly, even to a complete stranger, and therefore cannot involve a contextually determined newspaper. Note that the fact that example (145b) would be unintelligible in such a context shows that weak definites are sensitive to lexical restrictions; cf. Subsection B.

That weak definites do not have the regular reading of definite noun phrases is also clear from elliptical contexts such as (146). The two examples differ in that (146b) expresses that Jan and Marie have read the same, contextually determined, novel (the so-called strict reading) while (146a) does not presuppose that they read the same newspaper (the so-called sloppy reading). This follows on the assumption that weak definites differ from regular definites in that they do not refer to a unique referent in domain D.

Weak definites also differ from regular definites in that they are not compatible with a de re reading: example (147a) expresses that the speaker’s friends all read some, but not necessarily the same, newspaper (the de dicto reading), whereas example (147b) states that they all read the same, contextually determined novel (the de re reading). This again follows if we assume that weak definites differ from regular definites in that they do not refer to a unique referent in domain D.

Weak definites differ from regular definites in that they usually do not allow pluralization and modification; exceptions will be discussed in subsection B. This is clear from the fact that the examples in (148) do not allow the complex-predicate reading: these examples involve some contextually defined newspaper(s) and therefore cannot be used in out-of-the-blue contexts.

We have seen that some of the special properties of weak definites are related to the fact that they differ from regular definites in that they do not refer to specific entities in domain D: this explains why they cannot be the antecedent of a deictic pronoun, can be used in out-of-the-blue contexts, trigger sloppy-identity readings in elliptical contexts, and have a de dicto reading. We have also seen that they usually cannot be pluralized or modified. The inability of weak definites to be modified may be related to the fact that modification usually makes the definite noun phrase insufficiently “general” to be construed as part of a complex verbal predicate. This idea fits well with Guevara’s (2014) central claim that weak definites differ from regular definites in that they do not refer to entities, but to kinds. In this context, it is worth noting that definite generic noun phrases like de zebra in (149) share certain properties with weak definites; cf. Section 19.1.1.5. First of all, they do not refer to a specific entity in domain D, but to what we have called a prototypical member of the set denoted by the noun, and in addition they usually cannot be pluralized or modified; this is illustrated in (149b&c).

For further discussion of the similarity between weak definites and definite generic noun phrases, we refer the reader to Guevara (2014:§3). We will return to this issue briefly in Subsection C.

Weak definites are very common in the domain of art, entertainment, and recreation. The definite noun phrases in the sentences in (150) do not refer to a specific building or location, as can be seen from the fact that these sentences can be uttered out of the blue, but they do express that the speaker wants to be involved in some kind of activity, like visiting a museum or watching a movie outdoors.

In these cases, we are often dealing with more or less fixed combinations. This becomes clear in the constructions in (151), where we compare the acceptability judgments on noun phrases in their weak (non-referential) readings from more or less the same semantic domain (“performance art” in the broad sense). In (151) we see that noun phrases like het toneelthe play/theater and het concertthe concert cannot be used in constructions such as (150b). Note that speakers’ judgments seem to vary somewhat; the judgments in the examples below are our own.

151

Something similar can be observed in the examples in (152), which are quite common ways for the speaker to express that he wants to make his career in the world of film, theater, etc. It is not the case, however, that this applies to all artistic careers. For example, while it is possible for the speaker to use Ik wil bij de opera to express that he wants to pursue a career as an opera singer, he cannot express that he wants to become a member of an orchestra by saying Ik wil bij het orkest. This sentence can only be used in a referential reading of the noun phrase het orkest: “I want to join the orchestra”.

152

For the speaker to express that one would like to make a career as a musician (e.g. as a member of an orchestra), he could use the construction in (153a). This construction is a very restricted, idiomatic construction that does not allow any of the other definite noun phrases; (153b) is only acceptable in the literal sense, where het theater and, more marginally, de opera refer to buildings where performances take place (an option missing for de film).

Note that example (153a) involves a postpositional (i.e. directional) PP. To express that someone is a professional musician, we would use a prepositional (i.e. locational) PP instead, as in (154a). The (b)-examples show that this construction does not allow any of the other definite noun phrases either: only the literal meanings are available.

Weak definites are very common in sentences referring to outdoor activities, and they often specify the natural space in which the activity takes place. Importantly, however, they do not usually refer to specific locations, which is clear from the fact that all the sentences in (155) can be uttered out of the blue and followed by further specification of the location: Laten we naar ... gaanLets go to ...’, where the dots can be filled in by De Veluwe (nature reserve), Zandvoort (town on the coast) or FrankrijkFrance.

Again, there are idiosyncratic restrictions on the nouns that can head weak definites: the examples in (156) are not easily used spontaneously: the definite noun phrase het meer would normally be interpreted as referring to a specific location and the definite noun phrase de zee gives rise to a marked result, possibly because the bare noun zee is normally used in this case.

The examples in (157) show that weak definites can occasionally appear in plural form, which seems to contradict the claim in Subsection A6 that weak definites cannot be pluralized. However, when referring to a particular geographical area, duinen and bergen seem to be pluralia tantum, which is clear from the fact that their singular counterpart is marked or impossible in this context; hence they seem to be only apparent counterexamples to the claim.

Nouns denoting highly trained medical personnel like dokterdoctor, tandartsdentist, fysiotherapeutphysiotherapist and psychiaterpsychiatrist (but not verplegernurse) can be used as weak definites in cases like (158a). We are dealing with weak definites here, as is clear from the fact that such examples can be uttered out of the blue, e.g. as a reason for being late the next day, without the speaker knowing who will actually provide the medical care. That the definite noun phrase does not refer to some unique entity in domain D is especially clear in examples such as (158b), which can be used even when the speaker does not expect to speak to the doctor himself, but to a receptionist or a doctor’s assistant.

These examples show that the intended message can be quite different from the compositional meaning of the sentence: the examples in (158) express something like “I need some medical attention” or “I am going to see about some medical service”. Similar meaning extensions can be found when the noun denotes a medical institution, such as ziekenhuishospital in (159), which can be used to express that Jan needs to get/is getting specialist medical treatment, even if the speaker does not know which hospital will provide it.

Note in passing that the examples seem to contradict the claim in Subsection A6 that weak definites cannot be modified. The modified noun seems to denote a special subtype of hospital, which is apparently sufficiently “general” to license the complex-predicate meaning of “needing/receiving psychiatric help”.

The examples in (159) are interesting in that there is cross-linguistic variation on the question of whether the definite article must be expressed, as illustrated by the contrast between American and British English in (160), taken from Guevara (2014).

Weak definites and bare nouns usually seem to be able to perform the same function in PPs, the choice between them being completely idiosyncratic, as will be clear from the contrast between the two examples in (161); cf. P34.1, sub II, for more examples. This finding is important, because it underscores the claim that definite articles in weak definites do not have the same meaning as those in regular noun phrases; it actually suggests that they do have no indispensable semantic contribution to make.

Cases similar to those in (158) and (159) can also be found outside the medical domain, and are in fact quite common in relation to shopkeepers and shops, as shown in the examples in (162); note that in Dutch it is not always clear whether nouns like bakkerbaker, slagerbutcher and melkboermilkman refer to people with a specialized profession or to their shops.

Definite noun phrases are not necessarily interpreted specifically (or generically) in examples such as (163a): such examples are ambiguous and can mean either that the speaker takes a particular bus (e.g. the one that comes around the corner), or that he takes the bus qua means of transport (i.e. any bus). In many cases, such as (163b&c), the latter reading is clearly the preferred one.

On the second reading in (163a), de bus ‘the bus’ is not a uniquely referring noun phrase, but is interpreted as a subpart of the complex verbal predicate de bus nementake the bus, i.e. to engage in bus travel. Similar interpretations are possible for noun phrases with definite determiners with a variety of “means of transportation” as their head. To give an idea of the range of possibilities, some acceptable examples are given in (164a-c).

Note that the noun taxi in (164d) is special in that it requires an indefinite article. The fact that we apparently cannot predict which article will be used may simply reflect the more or less idiomatic character of this construction with the verb nemen (and its more colloquial near-equivalent pakkentake, fetch, catch). That we are dealing with idiom-like constructions may also be supported by the fact that there are clear idiomatic cases with this verb, which are given in (165b). Although (165a) seems to be a direct instantiation of the general pattern in (164), this example is also special in that the noun benenwagen can only be used in “means of transportation” contexts.

The examples in (166) show that the “means of transportation” interpretation of noun phrases with a definite article is also found in PPs in clauses with motion and position verbs: example (166a) involves an instrumental met-PP, while the PPs in (166b) are locational complementives.

The instrumental met-PP in (166a) sometimes alternates with a per-PP, with the Latinate preposition per, which systematically takes bare (i.e. determinerless) complement noun phrases. The fact that some of the cases in (166a) alternate with determiner-less cases in (167) supports the earlier suggestion that the definite article de in (166a) has no indispensable semantic contribution to make.

Like noun phrases denoting a means of transportation, noun phrases denoting a means of communication may contain a definite article that does not necessarily contribute to the notion of definiteness. This is particularly clear in (168b), where the article de can be omitted without any noticeable change in meaning.