The modifiers heel and half typically contribute the semantics of (moderately) high degree in the context of sometimes more or less fixed expressions of exaggeration. Metaphorical examples of this kind are given in (198). The degree modifiers heel and half are equally possible in (198), but the two differ in that the degree to which the verbal proposition holds is understood to be higher when heel is used than when half is used, but the difference is not very robust.

The degree reading is also obtainable in the non-metaphorical examples in (197), repeated here as (199). This is easiest with halve; example (199a) expresses on its degree reading that there were quite a large number of flowers scattered over the island, and (199b) that the village was hit by a wave of theft. Degree readings of this kind seem to be harder to get with heel.

Note that noun phrases with the degree modifiers heel/half require the main accent to be on the noun: een hele/halve volksverhuizing in (198a) and het halve dorp in (199b). Noun phrases with the quantifiers heel/half meaning “100/50%”, on the other hand, require the main accent to be placed on the quantifier: (197b) is realized as het halve dorp.

In sentences of the type in (200), the semantics of inflectible heel is also one of degree modification, which is best rendered by English quite. In contrast to the metaphorical high-degree cases in (198), inflectible heel in (200) does not alternate with half, but with degree modifiers like behoorlijk, flinkquite, generally with little or no difference in meaning.

The “quite” degree reading is impossible in definite noun phrases; examples such as (201a) are acceptable only on the totality interpretation of heel. Indefinite noun phrases with determiners other than the article een do not allow the “quite” degree reading either, as shown by the unacceptability of (201b&c). We conclude that the “quite” degree reading of inflectible heel depends on the presence of the indefinite article een.

Finally, note that inflectible heel can also modify the pronoun wat in (202a). Example (202b) shows that in this case heel also alternates with adjectival degree modifiers like behoorlijk and flinkquite. However, as can be seen in the English translations, one of the possible interpretations of heel wat is missing in the constructions with adjectival degree modifiers.

The interpretation of heel in binominal noun phrases of the type in (203) varies in a rather subtle way. Three interpretations are available for examples of this type. The high and “quite” degree interpretations of heel arise when the noun verzameling/lading is quantificational, whereas the adjectival meaning “complete” requires that the noun verzameling/lading is referential, i.e. assigned its literal meaning “collection/load”; cf. Subsection I for discussion.

Example (204) aims at bringing out the prosodic differences between the three interpretations of heel. The representations show that the two types of degree reading with the quantificational construal of verzameling require a single stress peak on the noun verzameling. The high-degree reading “a very large amount/number” in (204a) furthermore requires lengthening of the vowel, and the “quite” degree reading “quite a few” in (204b) requires an additional stress peak on the degree modifier heel. The referential reading of verzameling in (204b) is easily distinguished from the other uses, as the indefinite article can be replaced by other determiners such as the possessive pronoun zijnhis; the adjective heel receives main stress.

Note that the properties of the high-degree reading of inflectible heel in (204a) are also salient in metaphorical high-degree cases like those in (198): Hele genera—ties hebben dit lied meegezongen Whole generations have sung this song. Note also that the prosodic properties of the constructions in (204b&c) are preserved if we replace heel with a degree modifier like behoorlijk or an adjective like volledigcomplete. However, there is nothing that can replace heel in its high-degree reading with preservation of the intonation contour in (204a).

Examples of the type in (203) can be pluralized, but this results in the loss of two of the readings: only the high-degree interpretation seems to survive in (205). This is also clear from the intonation pattern; the main accent is on the nouns verzamelingen and ladingen, with lengthening of their stressed vowels. That the adjectival interpretation for hele in (205) is hard to get is also clear from the fact that the addition of e.g. a possessive pronoun to hele verzamelingen is pragmatically odd: #Ik heb mijn hele verzamelingen verkocht (lit.: I sold my whole collections).

The core lexical semantics of heel, i.e. totality, may not be completely absent in these high degree cases, as suggested by the fact that heel can be rendered in English with whole or entire. For the “quite” degree reading of heel, on the other hand, no translation with English whole or entire is possible in the general case; cf. example (200). This interpretation of heel thus seems far removed from the core quantificational semantics of this element.

The examples in (206) show that a degree interpretation is also possible when heel is used as an adverbial modifying an attributive or predicative adjective; cf. Section A26.1.2. Whether heel receives a high or a “quite” degree interpretation seems to depend on the adjective with which it is construed. Adverbial heel cannot be replaced by half in (206); degree modification of adjectives by half is possible only if half and the adjective form a compound (cf. halfzachte/*halve zachte drophalf-soft licorice), and this is not possible with the adjectives in (206).

It should be emphasized that the degree readings of heel do not depend on its construal as an adverb. That heel is not an adverb in the examples discussed up to (205) is evident from the fact, illustrated in (207), that it must inflect in accordance with the gender and number features of the head noun, whereas schwa-inflection is always optional with the adverbial phrases in (206).

We conclude the discussion of the degree readings of heel by addressing some additional types of examples classifiable under the “degree” header that typically involve predicatively used noun phrases. We begin with “quite” degree readings in clauses containing alalready and nog (best)actually. The examples in (208) differ from those discussed so far in that the adverb alalready must be present. Examples like (208a&b) are typically addressed to little boys or girls who are assumed to take pride in appearing older and wiser; the “quite” degree resides in the extent to which adulthood has already been “reached” or mimicked by the child in question. Especially in mildly ironic contexts, this type can also be used with nouns other than those illustrated in (208a&b), as a sentence like (208c) shows.

Since sentences of the type in (208) are typically used as statements directed to an addressee (especially little children), they usually have second-person pronouns as subjects. They are typically used as exclamations (as indicated by the punctuation used), but similar constructions can also be found in (tagged) rhetorical questions. An example is given in (209); note that in this context the adverb al, required in (208), is typically absent.

In (210) we find a dependency between noun phrases with heel on its degree reading and the adverbial nogstill/yet similar to that found in (208). Nog is often preceded or followed by the form best, the closest English paraphrase of which is something like “actually” of “frankly”. The two word orders seem to be semantically equivalent.

One might wonder what is the structural position of the adverbs al and best in (208) and (210). To investigate this, we will consider the topicalization constructions in (211), although it should be noted that the judgments are not entirely clear and are likely to vary between speakers. The examples in (211) show that it is difficult to strand the adverbs al and best when topicalizing the noun phrases with heel.

The degradation induced by stranding of the adverbs may indicate that the adverbs are subparts of the noun phrases, although pied piping of the adverbs does not yield very good results either.

An alternative would be to assume that al and best are independent constituents, which would account for the degraded status of the pied-piping cases in (212), and to say that the dependency between heel and these adverbs is similar to that between negative polarity items and their licensers; the degraded status of stranding al and best in (211) might then follow from the fact that the topicalized noun phrase with heel is outside the licensing domain of the adverb. We leave it to future research to decide whether this proposal has merit.

The examples in (213) are syntactically similar to those in (208a&b) and partially overlap in the lexical nouns heading the heel phrases (vent, heer, dame), but differ in that in (213) there is no adverb like alalready. Furthermore, the degree reading introduced by heel is that of high degree, directed towards the implicit qualities of the head noun; the nominal predicate een hele N attributes a high degree of excellence to the subject of the clause, which is noteworthy given that the nouns vent and kerel in (213a) often carry negative evaluative connotations.

Note that the connotation of excellence implicit in the nouns used in the examples in (213) is apparently missing in examples like ??Hij is een hele jongen/man (lit.: he is a whole boy/man) and ??Zij is een hele vrouw (lit.: she is a whole woman), since they do not yield the high-degree interpretation found in (213).

As its name suggests, negative-polar heel phrases must be licensed by some form of sentential negation. It differs from ordinary negative polarity phrases of the form ook maar Xany X in that it cannot be licensed in questions and conditional clauses; cf. Hoeksema (2007). This is illustrated in (216); note that we have systematically italicized the polarity items in the examples in this subsection.

In negative contexts, the licensing conditions of negative-polar heel phrases also differ from negative polarity items of the form ook maar X. The examples in (217), where the licensers are underlined, show that ook maar iemandanyone can be licensed in two ways: the licensing negation can be syntactically expressed (i) on a c-commanding constituent, such as the subject niemandnobody in the (a)-examples, or (ii) by the negative adverb nietnot in the (b)-examples. The two cases differ in that the adverb niet can only license the occurrence of ook maar iemand in some lower clause, whereas negation on a clausal constituent can also license its occurrence in its own minimal clause; this is illustrated by the acceptability contrast between the two primeless examples in (217). We refer the reader to Broekhuis & Klooster (2010) for a discussion of the syntactic distribution of ook maar X.

If we compare the examples in (217) with those in (218), we see that the pattern with negative-polar heel phrases is quite different: polar heel can be licensed by the negative adverb nietnot, but not by a c-commanding constituent: negative-polar heel must occur in the minimal clause of the adverb niet. All in all, this leads to a distribution of negative-polar heel phrases that is the inverse of ook maar X.

However, it is not the case that licensing by a clausemate other than the negative adverb nietnot is completely excluded; example (219a) shows that negative-polar heel phrases can also be licensed by clause-medial frequency adverbials such as nooitnever or nog nooitnot ... so far. Example (219b) further shows that such adverbials are like the adverb niet in that they cannot license a negative-polar heel phrase in a lower clause. For completeness, note that replacing the object in (219b) with ook maar iemand would lead to an acceptable result.

The embedded polarity items in the primed examples of (217) and (218) function as the objects of the embedded clause, but the examples in (220) show that we find the same contrast when they function as the subject; negative-polar heel phrases cannot occur as the subject of the embedded clause.

We can conclude from the discussion so far that negative-polar heel phrases differ from negative-polar ook maar X phrases in that negation cannot license a negative-polar heel phrase embedded in a lower clause (but see Subsection 4 for a possible counterexample). In fact, the restriction is even stricter: the negative licenser and the negative-polar heel phrase must be clausemates, i.e. they must be clausal constituents of the same clause. This explains not only the unacceptability of negative-polar heel phrases in object clauses in the earlier examples, but also the unacceptability of examples such as (221b), where the negative-polar heel phrase is part of a postnominal possessive van-PP of the object. We will return to this example in Subsection 3, where we will see that (221b) contrasts sharply with cases with a prenominal possessor.

The examples in (222a-b) show that arguments of om te-infinitives and (transparent) te-infinitives should be assigned a different status with respect to the clausal constituents of their matrix clause; while the former clearly cannot be seen as clausemates of the negation in the matrix clause (i.e. cannot be a negative-polar heel phrase licensed by it), the latter can. Example (222c) shows that arguments of bare infinitival complements of e.g. deontic modal verbs such as kunnento be able can also be seen as clausemates of the matrix negation. This fits well with the general patterns established for such complement clauses in Section V5.2.

Another difference between the licensing restrictions on negative-polar heel phrases and negative-polar ook maar X phrases is illustrated in (223): while polar heel can be licensed by implicitly negative verbs such as vergetento forget/to not know anymore, ook maar X cannot (although there are other negative polarity items that resemble heel in this respect; cf. Klooster 1993).

The potential of implicit negation for the licensing of negative-polar heel phrases is also clear from the acceptability contrast between the examples in (224): while the verb passerento pass plausibly features implicit negation in its lexical semantics (“not to be behind X anymore”), this is certainly not the case for the verbs in (224b).

Now that we have discussed some differences between negative-polar heel phrases and more common negative polarity items such as ook maar iemandanyone pertaining to the licensing negation, we can turn to discussing the syntactic functions of the former.

The examples in (225) show that there are various restrictions on the syntactic use of negative-polar heel phrases; although they can easily be used as direct or prepositional objects, they cannot be used as subjects.

Although acceptability judgments are not always clear and may vary from speaker to speaker, it seems that there is no general prohibition against using negative-polar heel phrases as subjects. The examples in (226) first show that negative-polar heel phrases cannot function as subjects of transitive and intransitive verbs.

However, the examples in (227) show that a negative-polar heel phrase gives rise to a much better result when it is a derived DO-subject: the active and passive constructions seem to have more or less the same status.

The relatively high degree of acceptability of the passive example in (227b) leads us to expect that subjects of unaccusative verbs can also appear as negative-polar heel phrases: this expectation is borne out, as shown in (228) for the monadic unaccusative verb vertrekkento leave and the nom-dat (i.e. dyadic unaccusative) verb bekorento please.

From the assumption that heel is a negative polarity item, the unacceptability of the examples in (226) follows directly, since negative polarity items never occur as underlying subjects of main clauses; cf. *Ook maar iemand heeft niet gelachen. However, the acceptability judgments on (228) show again that the licensing conditions on negative-polar heel phrases are different from those on negative polarity items such as ook maar iemandanyone, which cannot easily be used as the subject of unaccusative constructions in main clauses; cf. *Gisteren was ook maar iemand nog niet eens vertrokken and *Ook maar iets kan me echt niet bekoren. This is less clear for passive constructions, as examples of the type Er werd door niemand ook maar ietsSubj gezegdno one said anything at all sound much better and can occasionally be found on the internet.

Ditransitive constructions deserve special attention. A direct object can be realized as a negative-polar heel phrase but this leads to a degraded result in the case of an indirect object, as shown in (229) (but see Subsection 4 for a possible counterexample).

We will assume that (229) is the more general pattern, despite the fact that it seems to be at odds with the observation in the 2012 edition of this work that examples like (230a) are acceptable. It is not entirely clear what causes this difference in acceptability. One plausible account for the unexpected acceptability of (230a) appeals to the fact that the verb geven is a semantically light verb, so that een hand geven actually functions as a complex (i.e. transitive) predicate; cf. the nearly synonymous (b)-examples. Another type of account might appeal to the fact that the direct object follows negation in (230a&b); cf. Subsection 3 below. We leave this issue for future research.

We are not aware of other syntactic uses of negative-polar heel phrases. For instance, the negative-polar heel phrase in example (231a) is embedded in locational PP, which functions as a complementive selected by the verb zittento sit, and negative polar heel cannot be used. In (231b) we are not dealing with negative-polar heel, but with quantificational heel, as can be seen from the fact that the interpretation of the temporal adverbial phrase die hele nachtthat whole night does not change when we omit the negative adverb niet; we are dealing with a purely adjectival or quantificational instantiation of heel.

The previous section has shown that negative-polar heel phrases are prototypically direct/prepositional objects of the main verb licensed by sentential negation, expressed by the negative adverb niet. The examples in (232) also show that these objects must precede the adverb.

That the direct object must precede nietnot is not surprising, since example (233a) shows that this is generally true for definite objects in neutral (i.e. non-contrastive) contexts; cf. also Section 22.1.3, sub IV. However, this restriction is not expected for PP-objects, since (233b) shows that these can usually remain left-adjacent to the verb(s) in clause-final position. In fact, leftward movement of complement PPs is only possible if the nominal part of the PP is accented; cf. Section V13.2, sub IIIA.

The acceptability contrast between the (b)-examples in (232) and (233) can be seen as evidence for the claim that leftward movement of negative-polar heel phrases is an integral part of their licensing conditions. Den Dikken (2002/2006), for instance, claims more specifically that such phrases must undergo A'-scrambling into the specifier position of NegP (cf. Section V13.2 on Neg-movement). This seems to be a major improvement over the suggestion in the 2012 edition of this work that we are dealing with A-scrambling, since A-scrambling only affects nominal arguments and therefore cannot account for the obligatoriness of leftward PP-movement in (232b). Another reason for rejecting the earlier A-scrambling hypothesis is that this movement typically crosses the clause-medial adverbials, such as the modal waarschijnlijkprobably; the examples in (234) show that negative-polar heel phrases can occur between the modal verbs and negation (although they can also precede the modals).

The hypothesis that the negative-polar heel phrase must undergo A'-scrambling into a position preceding its licensing negation also accounts for the clausemate restriction discussed in Subsection 1. The relevant examples illustrating this, repeated in (235), are not acceptable because A'-movement of the subject/object did not apply; moreover, they cannot be saved by A'-scrambling of the negative-polar heel phrase out of the embedded clause across the licensing negation in the matrix clause, because finite clauses are islands for this kind of extraction.

The hypothesis also accounts for the fact, alluded to in Subsection 1, that there is a difference between negative-polar heel phrases that function as pre and postnominal possessors. The relevant examples are in (236); apparently the prenominal possessor turns the whole object into a negative-polar heel phrase (eligible for A'-scrambling), while the postnominal possessor does not.

This phenomenon mirrors what we see in (237), where the prenominal wh-possessor wienswhose makes the entire object eligible for wh-movement (i.e. the prototypical case A'-movement), while the postnominal possessor van wie does not.

It seems that we can conclude that negative-polar heel phrases must be A'-scrambled across the negative adverb niet in order to be licensed. Note, however, that many speakers at least marginally accept PP-complements with a negative-polar heel phrase in extraposed position; cf. (238).

The acceptability of (238) need not be taken to go against Den Dikken’s (2002/2006) claim that leftward A'-scrambling is obligatory, since there are theories in which extraposition is derived by leftward movement of the extraposed element followed by leftward movement of a (remnant) VP; cf. Den Dikken (1995), Barbiers (1995), and Kayne (2004) for a number of seminal proposals, and Hoeksema (2007:172-3) for a potential problem.

Finally, it should be noted that eligibility for A'-movement may not be sufficient for licensing negative-polar heel phrases, since it seems to be the case that the nominal part of PP-complements of adjectives like trotsproud and verliefdin love can be a negative noun phrase, but not a negative-polar heel phrase. Since it is generally assumed that the PP in (239a) is moved into the specifier position of NegP (cf. Section V13.2), there is no obvious reason to deny the negative-polar heel phrase the option of being A'-moved out of the AP in (239b).

We leave the unexpected acceptability contrast between the two examples in (239) to future research; we cannot exclude the possibility that we are in fact dealing with speaker variation, as Hoeksema (2007:fn.5) presents examples of the kind in (239b) with the PP in extraposed position as acceptable; cf. %dat Marie niet trots/verliefd is op die hele jongen.

Subsection 3 has shown that in the general case negative-polar heel phrases can be used as direct objects, but not as indirect objects; cf. (229). For the sake of variation, the examples in (240a&b) provide similar cases with the inherently negative verb onthoudento deny, which expresses a similar meaning to niet gevennot to give in (229). However, example (240c) shows that the indirect object can take the form of a negative-polar heel phrase, provided that the direct object takes the same form. Note that elements that indirectly license an otherwise unacceptable negative-polar heel phrase are underlined (because they act as licensers) and italicized (because they are negative polarity items themselves).

Given the complexity of the construction, we provide another set of examples in (241), all of which are perfectly acceptable without polar heel; cf. Hij wil die student die constructie niet uitleggen. Den Dikken (2002) concluded from examples like these that the licensing of negative-polar heel in the indirect object is “parasitic” on the licensing of negative-polar heel in the direct object.

The form of parasitic licensing of negative-polar heel phrases in the (c)-examples is not an isolated case, but can also be observed in embedded clauses. Subsection 1 has shown that while negative-polar heel phrases can be licensed by a negative element in their own clause, they cannot be licensed by a negative element in a higher clause; the examples illustrating the latter are repeated here as (242).

However, these examples can be improved considerably if the embedded clause contains another polarity item. We start with the subject case in (242a): the negative polarity phrase dat hele meisjethat whole girl seems to be licensed by the presence of the negative polarity items ook maar iemandanyone in (243a) and ooitever in (243b), which in turn are licensed by the negative adverb niet in the matrix clause. The percentage signs indicate that some speakers do not easily accept examples of this kind (perhaps because of their complexity); we will ignore this issue for the rest of this subsection.

Note that the parasitically licensed negative-polar heel phrase in (243) resembles its directly licensed counterpart in that it must precede its (indirect) licenser, but differs from it in that it need not be c-commanded by its licenser; the subject dat hele meisje is in the regular subject position, which c-commands the direct object ook maar iemand as well as the clause-medial frequency adverb ooit.

Let us now consider the case with a direct object in (242b). Example (244a) first shows that the negative-polar object dat hele meisje cannot be licensed by an ook maar X phrase functioning as subject. The acceptability contrast between (243a) and (244a) suggests that the negative-polar heel phrase must precede (i.e. c-command) the negative polarity item that indirectly licenses it: a negative-polar heel phrase functioning as subject can be licensed by a regular negative polarity item functioning as direct object, while a negative-polar heel phrase functioning as direct object cannot be licensed by a regular negative polarity item functioning as subject. That precedence is indeed the relevant factor is supported by the fact, illustrated in (244b), that a negative-polar heel phrase functioning as direct object can be licensed indirectly if it precedes the negative polarity item ooit, which is licensed directly by the sentential negation in the matrix clause.

Finally, example (245) shows that indirect objects can also be indirectly licensed by the negative polarity item ooit, provided that they precede it. Note that the position of the direct object does not affect the acceptability judgments, i.e. it can intervene between the parasitically licensed negative-polar heel phrase and its licenser ooit.

There are at least three additional restrictions on parasitic licensing of negative-polar heel phrases. The first restriction is that the negative-polar heel phrase and the polarity item that indirectly licenses it must be clausemates. This is illustrated by the examples in (246): because (246a) shows that the negative-polarity item ook maar iemandanyone can be licensed by the sentence adverb over an (in principle) unbounded distance, the unacceptability of (246b) must be due to the fact that the negative-polar heel phrase and ook maar iemand are not part of the same clause.

The second restriction is that the object clause containing the parasitically licensed negative-polar heel phrase and the negative polarity item that indirectly licenses it must itself be a clausemate of the direct licenser of the negative polarity item. This is illustrated by the unacceptability of the examples in (247): the (a)-example should be compared to the examples in (243), while the (b)-example can be compared to (244b). The percentage sign in (247b) is used because Den Dikken (2006:§2.1) considers this example grammatical, while recognizing that it is not fully acceptable and even rejected by many speakers; cf. also Hoeksema (2007:§1). We leave the status of (247b) open for further discussion.

A third restriction is that the clause containing the parasitically licensed negative-polar heel phrase must be c-commanded by negation. While both the active and passive constructions in (248a&b) with the inherently negative verb ontkennento deny seem acceptable, the passive example in (248c) is degraded; the unacceptability of (248c) has been attributed to the fact that the subject clause is in the regular subject position in this case.

Finally, we can observe that parasitic licensing is restricted to argument clauses. Example (249a) shows that the negative-polarity item ook maar een beetje can be licensed by the sentential negation expressed on the antecedent of the relative clause containing it. However, parasitic licensing of the negative-polar heel phrase in (249b) by this negative-polarity item seems impossible.

We want to conclude our discussion of parasitic licensing with two findings from Hoeksema (2007). First, he provides more examples of polarity items that are covered by the clausemate restriction and can be licensed parasitically. One example is meeranymore in (250): the contrast between the primeless and primed examples first shows that negative-polar meer can be licensed by sentential negation, provided that it is expressed by one of its clausemates.

The examples in (251) further show that the use of negative-polar meer in the primed examples in (250) improves considerably with the addition of the negative polarity item ooitever. We added the percentage sign to indicate that some speakers prefer the non-polar item nog instead of meer in this context, but as Hoeksema himself points out, this does not affect the point made here that meer is better in combination with ooitever than it is by itself. Note that the parasitically licensed polarity item meer differs from parasitically licensed negative-polar heel phrases in that it need not (in fact: cannot) precede its licenser ooit, which shows that the precedence restriction established for parasitically licensed negative-polar heel phrases does not carry over to this case; meer apparently does not have to undergo A'-scrambling.

Second, Hoeksema shows that parasitic licensing of negative-polar heel phrases is not restricted to negative contexts. The (a)-examples in (216) have shown that negative-polar heel phrases and negative polarity items of the form ook maar X differ in that only the latter may occur in questions; this is again illustrated by the unacceptability of (252a). Example (252b) shows that the acceptability improves considerably if we add the negative-polarity item ooitever (which occurs freely in questions; cf. Heb je hem ooit gezien? Have you ever seen him?).

The (b)-examples in (216) have also shown that negative-polar heel phrases differ from regular negative polarity items such as ooit and ook maar X in that they cannot occur in conditional contexts; this is again illustrated by the unacceptability of (253a&b). The primed example shows that acceptability increases with the addition of a negative polarity item like ooitever or ook maar X any X.

As in the examples in (243), the parasitically licensed negative-polar heel phrase in the primed (b)-examples resembles its directly licensed counterpart in that it must precede its (indirect) licenser. However, it differs in that it does not need to be c-commanded by its licenser: the subject dat hele meisje is in the regular subject position, which c-commands the direct object ook maar iemand as well as the clause-medial frequency adverb ooit.

An important connotation of the negative-polar heel construction is condescension: Bert Mulder and Marie are clearly not held in high esteem by the speakers in (254). Consistent with this is the fact, also illustrated in (254), that polar heel is typically used in combination with the distal demonstrative die/dat, which can also be used to express a negative evaluation on the part of the speaker; cf. Section 19.2.3.2, sub IID.

The proximate demonstratives deze and dit, on the other hand, generally produce an awkward result in this context and are ruled out altogether if the head noun is a proper noun; cf. the examples in (255).

Negative-polar heel constructions are also sometimes possible with definite articles (especially neuter het) when the head noun is inherently disparaging, like wijfbitch and neuter mens(female) creature. Note, however, that the forms in (256) marked with a percentage sign do occur on the internet.

The examples in (257) show that the use of the indefinite article eena leads to unacceptability. However, the unacceptability of these examples need not be related to the presence of negative-polar heel, since they are equally unacceptable without heel; non-specific indefinite noun phrases simply cannot move across the negative adverb niet.

The direct object would have to follow niet to satisfy this movement restriction, but, as discussed in Section 19.1.5, the sequence of the clausal negation niet and the indefinite article een is generally avoided by using the determiner geenno instead, as in (258). These examples, to the extent that they are acceptable in any context at all, certainly lack the specialized semantics of negative-polar heel.

We can conclude from this that negative-polar heel cannot be used in indefinite noun phrases headed by the articles een and geen, but it is still not clear whether this is due to a co-occurrence restriction between these determiners and negative-polar heel, or some other reason. We have already seen that the unacceptability of the examples in (257) is probably due to a more general restriction on leftward movement of non-specific indefinite noun phrases. The unacceptability of the examples in (258) may be due to the restriction that negative-polar heel phrases c-command their licenser; cf. Subsection 1. However, that we are dealing with a co-occurrence restriction is suggested by the fact that negative-polar heel is also blocked in indefinite noun phrases with the determiner zo’nsuch a: the primeless examples in (259) show that such noun phrases can be moved across negation, while the primed examples show that negative-polar heel is excluded in both surface positions.

Finally, the primeless examples in (260) show that combining the negative polarity use of heel with prenominal possessors gives rise to a marked result; fully acceptable possessive examples can be obtained by placing the possessor in postnominal position, as in the primed examples.