Table 4 has already shown that predeterminer alle + Num can precede the definite article de and the distal demonstrative die with equal ease; the relevant examples are repeated here as (45a&b) along with their counterparts with allebeiboth. However, the proximate demonstrative deze in (45c) seems to be rather marked in a position linearly following the predeterminer alle + Num and allebei. In this respect, predeterminer alle differs from al, which is perfect to the left of the proximate demonstrative (cf. al deze mannen/huizenall these men/houses), and instead resembles predeterminer heelall/whole; cf. Section 21.2.2.1, sub I.

The marked proximate demonstrative cases improve in contexts of the type in (46a&b), where the demonstrative receives contrastive accent (although some speakers may still consider such cases slightly contrived). Examples of this type become fully acceptable to all speakers under backward conjunction reduction, as in the primed examples. In contrast, NP-ellipsis in the second conjunct leads to unacceptability, as shown in the doubly-primed examples; this is entirely due to the presence of the predeterminers alle tien and allebei, because the sentences are perfect without these elements. Replacing these predeterminers with bare al also leads to a slightly better result, as can be seen in Section 21.1.2.1, sub IIA.

The examples (47a&b) show that possessive pronouns can be placed between alle + Num/allebei and the projection of the noun, provided that they are prosodically weak; the use of contrastive stress on the possessive pronouns is impossible for at least some speakers (cf. %alle tien/allebei mijn boeken). The strong forms in (47b) are marked to varying degrees; in particular onzeour and huntheir, which have no weak form, are relatively acceptable, but jullie, which can only be reduced under the strict conditions discussed in Section 19.2.2.3, seems completely unacceptable. The examples in (47c) show that (semi-)genitival possessors are usually acceptable to the same marginal degree as the strong possessive pronouns in (47b) when they have the form of a full noun phrase, but for some reason completely impossible when they have the form of a proper noun, as in (47d). Note that the examples in (47) are more or less acceptable if all substitutes for alle + Num; cf. Section 21.1.2.1, sub IIA for more examples.

The alle + Num example in (47a) alternates with a construction discussed in Section 21.1.2.1, sub IIC, in which bare al precedes and the cardinal follows the possessor of the noun; cf. al m’n tien boekenal my ten books. This alternation does not seem semantically innocuous, the judgments may be subtle. Consider the pair in (48), used in a context in which a selection committee is finalizing a shortlist, and each of the committee members has drawn up a list of three top candidates. In this context, alle drie onze kandidaten favors a reading in which all committee members happened to select the same candidates (resulting in a shortlist of also three candidates), while al onze drie kandidaten seems more felicitously used in a situation in which some committee members selected different candidates (resulting in a shortlist of more than three candidates).

Finally, the examples in (49) show that alle + Num and allebei can also be used with nominalized possessive pronouns. The article preceding the noun must be de, because alle + Num and allebei do not combine with singular noun phrases; this excludes the article het, which is only found with singular noun phrases.

Predeterminer alle + Num cannot be construed with noun phrases containing the indefinite article een in (50a), which is not surprising given that this article selects a singular count noun. For completeness’ sake, (50b) shows that this restriction also applies to the spurious article een (pronounced /ən/) found in exclamative constructions, which can be combined with a plural noun phrase in the absence of the predeterminer.

Predeterminer alle + Num can sometimes be combined with bare plurals, depending on the cardinal. The broad empirical generalization seems to be that with “high” cardinals, such as vijfentwintig, the article de is preferably omitted (although some speakers find this example perfectly acceptable with it), while with “low” cardinals the pattern without the determiner de is awkward, to an extent that varies somewhat from speaker to speaker and from case to case; cf. Perridon (1997:184). Postma (2015/to appear) claims on the basis of a limited survey that with twee, drie and vier, as in (51a&b), the use of the article de is practically obligatory; the same seems to hold for allebei. With the cardinals vijffive to twaalftwelve in (51c), the use of the article de is essentially optional, although there seems to be an additional phonological restriction at play in that the article de must be preceded by a stressed syllable: the informants considered alle zeven/negen boekenall seven/nine books, in which the cardinals end in a schwa, to be degraded. All other cases are disfavored or rejected by the informants, which Postma attributes to the fact that they are morphologically complex cardinals; cf. der-tien and vijf-en-[twin-tig] in (51d). Postma’s results agree well with our findings reported in the first edition of SoD, which were based on a rough internet search. For completeness’ sake, note that the noun phrases in (51) are all interpreted as definite.

Finally, note that when the cardinal is modified by an approximative adverb, combining alle + Num with a bare plural is the only possibility; adding the article de leads to an unacceptable result even with relatively “low” cardinals like tien; cf. alle tien (de) deelnemers.

Example (53a) shows that the indefinite determiner-like elements dat/dit soortsuch, discussed in Section 18.1.2, are not compatible with alle + Num/allebei to their left. Example (53b) shows that the same holds for pseudo-partitive noun phrases, discussed in Section 20.4, sub I, with the indefinite determiner-like element van diesuch. The noun phrases in (53) thus behave like indefinite noun phrases with the article een in (50a).

Since the predeterminers alle + Num and allebei already contain a cardinal expression, the addition of other quantificational elements is blocked; the examples below are all unacceptable. Cardinal numerals are also excluded, of course, since this would lead either to tautological redundancy or to a contradiction: *alle drie de drie/vier boekenall three the three/four books.

Alle + Num and allebei can form a constituent with a plural personal pronoun, provided that the pronoun is realized in its strong, unreduced form and that it precedes the quantifier. That the string zij alle tien/allebei in (55a&b) is a constituent is clear from the fact that it can precede the finite verb zijn in the main clauses (cf. the constituency test). Note that (55b), in which the pronoun has an inanimate referent, is much worse than (55a), in which the pronoun is animate, which is probably related to the fact that strong pronouns are typically interpreted as [+human]; cf. Section 19.2.1.1, sub V.

Note that the question marks in (55a&b) indicate that the use of the strong pronoun zijthey is marked compared to its use in the variants in (56), where the pronoun and quantifier do not form a constituent. This use of Alle + Num and allebei as floating quantifiers imposes no restriction on the form of the pronoun, i.e. the strong and weak forms in (56) are both perfect; cf. Section 21.1.4 for further discussion of the use of Alle + Num as a floating quantifier.

Since proper nouns are normally uniquely referring expressions, it is to be expected that they do not co-occur with predeterminer alle + Num, as shown in (57a). The exceptional case in (57b) is not a counterexample to this claim; the plural proper noun is preceded by the definite article de and therefore behaves like a count noun.

In colloquial language, simplex alle cannot be combined with the definite article de or the demonstratives die/dezethose/these (although the sequence alle de/die/deze can be found in archaic and very formal language). Beide differs from alle in that it can be used to the right, but not to the left, of these determiners. Note that the fact that beide is not in complementary distribution with the definite article indicates that beide cannot be treated as the result of a merger of the morpheme bei, which is also found in the predeterminer allebei (see Section 21.1.2.2), and the definite article.

There is no way to salvage the unacceptable examples in (64a) by contrastive accentuation; the bad cases remain bad regardless of the context in which they are used. On the other hand, the contrastive example in (65a), a contextualized variant of (64b) with beide, though somewhat clumsy, is structurally well-formed. Backward conjunction reduction is marginally possible in (65b) when applied to beide mannen, but highly awkward when applied to mannen alone. NP-ellipsis in the second conjunct strengthens this distinction; (65c) shows that it is fine with beide mannen elided, but straightforwardly unacceptable with just mannen undergoing ellipsis. -

Alle does not occur in noun phrases with a prenominal possessor. Beide, on the other hand, is again possible when it appears to the right of the possessor, as in (66b). The acceptability of these examples is unaffected by the complexity of the possessor; all variants of (66b) with beide in the right-hand slot are perfect, while all their counterparts with alle in (66a) are unacceptable.

The examples in (67a) show that neither alle nor beide combine with nominalized possessive pronouns. The unacceptability of (67a) is striking in view of the impeccability of (67b), which contains the predeterminer counterparts of alle and allebei. This contrast between (67a&b) is another indication that alle is not the result of a merger of the predeterminer al in (67b) with the following definite article. The difference between al and allebei in (67b') has to do with the fact that only the former can be combined with non-count nouns. That het mijne is necessarily singular excludes allebei; cf. the earlier discussion of example (49).

Alle and beide cannot be construed with noun phrases containing the indefinite article een, not even in the exclamative context in (68b), where the indefinite article can be combined with a plural noun phrase in the absence of the predeterminer.

Example (64b) has shown that in addition to beide mannenboth men, de beide mannenthe both men is also acceptable. Note, however, that the interpretation and concomitant syntactic distribution of the two forms are not identical. It seems that when beide is used without a determiner, it is semantically on a par with allebei + Det: allebei de mannenboth the men. Postdeterminer beide, on the other hand, is semantically distinct and behaves more like the cardinal twee: de beide/twee mannenthe two men. The difference between the pre and postdeterminer beide can be seen in the examples in (69). While the preferred interpretation of (69a) is that the two Koreas have made a joint decision, the interpretation for (69b) is that each of the two Koreas has made a decision on its own. In this respect, (69b) seems to be on a par with (69b'), although the latter also allows the interpretation that a joint decision was made.

That the two (b)-examples pattern together and are distinct from the pair in (69a) seems to be confirmed by (70): while the two cases in (70a) are perfectly acceptable, those in (70b) seem marked (although such cases can be found on the internet, especially with beide).

We conclude this subsection on determiners by paying more attention to the merger approach to alle, according to which it is a merged form of al and the definite determiner. The overall conclusion from the above discussion is that simplex alle combines only with bare plural noun phrases. The complementary distribution of alle and the determiners makes it possible to hypothesize that alle occurs in the determiner position, which would of course support the merger approach to alle. However, the examples in (71) provide evidence against this approach, since they show that alle does not behave like a definite determiner when it comes to the determination of adjectival inflection. While the definite article in (71c) triggers the inflectional -e ending in the attributive modifier of the noun, the attributive modifier in the alle phrase in the attested example in (71a), taken from Perridon (1997:182), inflects as in the indefinite noun phrase in (71b); the judgments given are our own.

Although alle does not qualify as a definite determiner (which in turn may indicate that it does not occupy the determiner position in the DP), the noun phrase in (71a) has the outward appearance of a definite/strong noun phrase in the sense that it cannot be used as the subject in the expletive er construction in (72b). Recall that this holds for all alle phrases (except those in which alle has a high-degree reading); cf. Section 21.1.1, sub IV.

Although the adjectival inflection suggests that the noun phrase alle slim geknoei is internally indefinite, the same noun phrase behaves like a definite/strong noun phrase in its external syntactic distribution. This paradox has received little or no attention in the literature; we will not attempt to resolve the tension between these seemingly irreconcilable characteristics of alle phrases, but we do want to emphasize our earlier conclusion that alle does not behave as a merged form of al and the definite article.

The indefinite determiner-like elements dat/dit soortsuch, discussed in Section 18.1.2, are not compatible with alle/beide, regardless of their position relative to the quantifiers. The same is true for the indefinite determiner-like element van diesuch (lit.: of those) in the pseudo-partitive construction.

Simplex alle and beide cannot co-occur with other quantifiers in a noun phrase, as shown by the fact that the examples in (74) are all unacceptable, regardless of the relative order of alle/beide and the additional quantifiers.

Word order patterns such as alle honderd boekenall hundred books do yield an acceptable result with higher cardinals, but alle does not directly quantify the noun phrase in such constructions, and arguably forms a constituent together with the cardinal; cf. the discussion in Subsection IB.

Another consideration concerning alle/beide that casts doubt on a merger approach is illustrated in (75); unlike the predeterminers al and allebei, alle and beide can be combined with animate personal pronouns into a single constituent. As in the case of alle + Num, the pronoun is realized in its strong form and precedes the quantifier; note that the quantifier is obligatorily adorned with a mute orthographic -n.

That the pronoun must be strong may explain why the examples in (76) are degraded on the intended reading, as pronouns referring to non-human entities usually have the weak form; cf. Section 19.2.1.1, sub V.

The examples in (75a) are marked compared to their variants in (77). These so-called floating quantifier constructions, which allow both the strong and the weak form of the pronoun, are discussed in Section 21.1.4.

Given the account of the degraded status of the examples in (76), the fact that the pronoun can be weak in the constructions in (77) correctly leads us to expect the examples with inanimate subjects in (78); note that the quantifier cannot be followed by the mute, orthographic –n.

Note also that the forms alle(n)/beide(n) belong to the formal lexicon: in everyday language these forms are usually replaced by the invariant forms allemaal and allebei, as in the examples of (79).

Although the personal pronouns and the floating quantifiers do not form a single constituent in (79), they do so in the examples in (75). In (75a) this is evident from the fact that the string zij allen/beiden occupies the position to the left of the finite verb (the constituency test). That the string hen allen/beiden in (75b) also forms a single constituent is clear from the fact that it functions as the complement of a preposition. The fact that it is possible for allen/beiden to quantify and form a single constituent with personal pronouns provides evidence that allen/beiden is not the result of contraction of al/bei and a definite article, since pronouns are never combined with articles in Dutch; cf. *de zij/hen/hun (lit.: the they/them/their).

The same is true for the quantified possessive pronouns in (80); the fact that aller and beider are marked with genitive case shows that we are dealing with constructions that are not part of the present-day language, but with relics from older stages of the language.

For completeness, note that the fact that (80b) has the form u instead of the possessive form uw suggests that the pronouns in these examples are not genitive forms, and the fact that (80c) has hun instead of hen further suggests that we are dealing with dative forms. If so, this would explain why the pronoun ons in (80a) is always uninflected, as in ons/*onze aller vriendthe friend of all of us, because unlike the possessive pronoun, the object pronoun ons does not have a schwa-inflected form; for more information, see onzetaal.nl/taalloket/uw-u-beider-aanwezigheid.

Finally, note that the independently used neuter singular demonstratives dit and dat can also be combined with al to form a single constituent; the form then used is alles. This combination is not possible with prenominal demonstratives; cf. *dit/dat alles boek (lit: this/that all book).