Abstract and Keywords
This chapter deals with the DP and its various layers. The lowest layer is the lexical domain, the NP. On top of the NP, there is a grammatical domain, calledsee Inflectional Phrase (IP), which contains the nominal inflectional categories of number and definiteness. The highest domain is the referential domain, the DP. The noun may be followed by complements and adjuncts, mainly in the form of PPs, and preceded by adjectives or quantifiers. Definiteness may be expressed as a preposed definite article or as a suffix on the noun. A non-modified noun moves to D, but an adjective blocks this movement and the definite article is spelt out as a separate word in D. There are several ways of expressing possession, especially in Norwegian, where the possessor can be either pre- or postnominal. In the other languages it is prenominal. Restrictive relative clauses are right-adjoined to IP, non-restrictive to DP. Universal quantifiers are generated above DP.
This chapter deals with the determiner phrase and its various layers. The lowest layer is the lexical domain, the noun phrase (NP). On top of the noun phrase, there is a grammatical domain, called inflectional phrase, which contains the nominal inflectional categories of number and definiteness. The highest domain is the referential domain, the determiner phrase (DP).
2.1 The noun phrase
The noun phrase is headed by a noun, which may be preceded and followed by various kinds of elements. There may be a specifier in the form of a possessive, and left-adjoined modifiers in the form of adjectives and quantifiers. Elements that follow the noun are of many different categories and semantic functions. They are either complements or adjuncts. Complements are postnominal phrases that bear some sort of argument relation or other close semantic relation to the head noun, and which do not normally allow other elements between them and the noun (except possessives). Adjuncts are right-adjoined to NP, and thus normally follow complements. The distinction between complements and adjuncts is by no means clear-cut, especially since the relative order of postnominal phrases may be rather free. Nor is this distinction descriptively important; it is mainly of theoretical concern, since there is room for only one complement in the nominal projection, while there is no principled limit to the number of adjoined phrases.
The structure of a noun phrase with a possessor and a complement is as in (1).
The Scandinavian noun has inherent gender. In standard Danish and Swedish nouns are either common gender or neuter. In Norwegian nynorsk and most Norwegian dialects there are three genders, masculine, feminine, and neuter. Norwegian bokmål comes in different varieties, allowing either a two-gender or a three-gender system. Count nouns are inflected for number, singular and plural. There is no case inflection in nouns in the standard languages. There is a possessive suffix –s, but this is no longer a case suffix, it is rather to be considered a phrasal clitic (184.108.40.206). Scandinavian has a definite article which is expressed as a suffix on the noun (see 2.3.1). This article agrees with the noun for gender and number.
The Scandinavian noun thus has four different forms:
Note that in Danish and Swedish the definite suffix is added directly on to the singular or plural indefinite form of the noun, like a clitic, while in Norwegian the singular stem or the plural suffix may be modified by the addition of the definite article, which then is more affix-like: N jente + a > jenta; jenter + ne > jentene.1
Nouns derived from transitive verbs may take a complement in the form of a PP with the preposition D af NS av ‘of’, whether they denote the action or the agent.
If the complement corresponds to a prepositional object (see 5.4.3), the same preposition is normally used with the noun as with the verb.
A complement corresponding to a recipient or an indirect object is introduced by the preposition DN til S till ‘to’.
Nouns derived from adjectives also take complements.
Certain nouns may have a verbal content without being lexically derived from verbs, and therefore still take complements with an object role.
With inalienable possession the possessor may be added with the preposition på ‘on’.
The prepositions D af NS av ‘of’ and på ‘on’ are used in partitive constructions.
A somewhat similar construction is (9).
Other types of postnominal phrases which could be considered complements by the definition given above, include partitive constructions with a noun followed by a mass noun or a plural noun.
Clausal complements are often introduced by prepositions, but in Swedish the preposition is frequently omitted.
Various kinds of phrases, with different semantic functions, may be right-adjoined to the noun.
A nominal or an adjective predicated of the head noun is added by means of the particle som ‘as’.
Even finite clauses may be adjoined to a noun.
These are also nouns with a verbal content, and the clauses correspond to adverbial clauses adjoined to VP; see 220.127.116.11.
The NP is merged as the complement of a grammatical head called I (for Inflection).2 This is the seat of grammatical features for number and definiteness. One important difference between Norwegian and Swedish on the one hand and Danish on the other, is that only the former two have a definiteness feature in I. In Norwegian and Swedish the noun moves to I to receive the nominal features of number and definiteness (in Danish number only). Through this operation the noun precedes any possessive element. This is seen in Norwegian, where possessives may follow the noun on the surface, while preceding a complement.
In Danish and Swedish, and under certain circumstances also in Norwegian, the possessive determiner raises to D, and thus again precedes the noun. The possessor will be further discussed in section 2.3.4. The structure of (18Nb) is as in (19).
Prenominal modifiers are adjectives or quantifiers, which are left-adjoined to IP,3 quantifiers above adjectives.
Modifiers are phrases, not heads, as can be seen from the fact that they themselves can be modified, (2D,Nn), and also from the fact that they can be coordinated, (2S).
Adjectives agree with the noun they modify for gender, number, and definiteness. The definite form ends in DN –e and S –a/e, and has no gender or number distinction.4
Multiple adjectives may occur in front of the noun also without being coordinated. They are then adjoined one on top of the other.
Participles function as modifiers in the same way as adjectives do.
The genitive of measure words, as well as of certain other more or less fixed expressions, can function as modifiers.
If the adjective has a complement following it, the whole adjective phrase has to follow the noun.
On the other hand, the adjective may precede the noun, while its complement follows it. This is first of all found with degree words, which include not only adjectives in the comparative, but also words like ‘same’, ‘such’, ‘other’, etc. (more on comparison in 8.3.5).
Participles used as modifiers also follow the noun when they have their own complements.
On the other hand, the adjective may be preceded by a prepositional phrase which further modifies the adjective.
This construction is rather marginal in Norwegian, used only in formal style bokmål, but more common in written Danish and Swedish (see 3.2.2).
Adjectives may be used by themselves without a head noun, either in a general sense, or because the noun is omitted under identity with a previous noun.
Existential quantifiers and a few other quantifier-like elements are left-adjoined above the adjective, and thus precede the noun and any adjectival modifier on the surface, as in (1). Universal quantifiers, which have a different distribution, will be treated in section 2.6.1.
The quantifiers occurring in this position are the existential quantifiers DN mange S många ‘many’, få ‘few’, D meget Nb mye Nn mykje S mycket ‘much’, D lidt SN lite ‘little’, D nogen Nb noen Nn nokon S någon ‘some, any’, ingen ‘no’, the cardinal numerals, and a few other quantifcational words, such as DN nok S nog ‘enough’. The ordinals are adjectives.
Like adjectives, quantifiers agree with the head noun for number and gender. The words for ‘much’ and ‘little’ are used with mass nouns, as in (1S), and they have singular forms only. The singular forms of the others are also mostly used with mass nouns.
The word DNbS en Nn ein ‘one’ with a phonologically reduced pronunciation serves as the indefinite article. Examples are seen in (3–5) above. As the indefinite article is a quantifier, and not a determiner, it can be used with non-referring predicate complements.
This use of the indefinite article is further discussed in section 18.104.22.168.
The word en/ein ‘one’ preceded by a determiner has the definite inflection (see 2.3.1).
With a certain evaluative or exclamative meaning, the adjective is preceded by så ‘so’, and followed by the indefinite article.
The plural forms D nogle Nb noen Nn nokre S några ‘some’ serve as the plural of the indefinite article.
A quantifier preceding an adjective may be repeated after the adjective as a replacement of a missing head noun.
Quantifiers may be modified by words like DNb bare Nn berre S bara ‘only’ and D næsten N nesten S nästan ‘almost’.
The regular order of quantifiers and adjectives predicts that cardinal numerals, which are quantifiers by definition, will precede ordinal numerals, which are adjectives, superlative adjectives, and adjectives with a ‘superlative’ meaning.
The ‘superlative’ adjectives all have the definite suffix, as expected for adjectives following a definite determiner.
This may not be an exception to the general rule, however. In the phrases in (20), the adjective in –ste may be analyzed as a modifier of the numeral, whereas in (19), the adjective modifies the noun. The structure of (19D), then, is as in (21a), and the structure of (20D) is as in (21b).
Quantifiers are used by themselves without a head noun.
2.3 The determiner phrase
The highest nominal projection is the DP, where the D head is the seat of the definiteness features and the referential properties of the nominal. In a definite DP, the D head has to be visible, which means that it must be phonologically filled. Definiteness can be overtly marked by a definite article, a demonstrative, or a possessive element. Proper names are inherently definite, and do normally not take a definite article.
2.3.1 The definite article
The definite article has two forms in Scandinavian, either as a suffix on the noun, or as an independent word. The suffixed article is added to the noun through (p.21) N-movement to a functional head position containing the definiteness marker. In Norwegian and Swedish the noun joins with the definiteness suffix in I (see 2.1.4). This suffix agrees with the noun for gender and number, and is added after the number inflection.5
When the definite noun is not preceded by a modifier, it moves to D, where it lexicalizes the definiteness feature.
In Danish, the definiteness feature is not merged in I, only in D. When not preceded by a modifier, the noun moves to D and joins with the definiteness marker.
So far, the result is the same as in Norwegian and Swedish: a noun with a definite suffix, the difference being the position of the feature [+DEF].
When the noun is preceded by a modifier—adjective or quantifier—the movement to D is blocked.6 The definite noun stays in I, and the definiteness feature is lexicalized by the independent definite article den ‘the’. The article agrees with the noun for gender and number, and the adjective has the definite adjectival ending –e/a.7
This construction is known as ‘double definiteness’, since the noun has the definite suffix in addition to the prenominal independent article.
Danish does not have double definiteness. There is no definiteness feature in I, so when the noun is blocked from movement to D, it does not receive a definiteness suffix.
In Swedish, the preposed definite article may be omitted, while the noun still has the definite form. This happens especially with adjectives in the superlative, with ordinal numbers, or with the superlative-like första ‘first’, sista ‘last’, nästa ‘next’, enda ‘only’, etc.
Even adjectives in the positive may be used without the article if they have a restrictive reference, if the noun has a generic or non-specific reference, and if the two make up a more or less fixed expression.
This is less common in Norwegian, but it is found in some fixed expressions, especially with negation.
The definite determiner phrase in the last example can be considered a negative polarity item, and the sentence would be ungrammatical without the negation.
Indefinite nouns in definite phrases are used in connection with absolute superlatives even in Norwegian and Swedish.
In Norwegian bokmål there are a few lexicalized phrases with an indefinite noun used in definite DPs. They are generic expressions or names of organizations or institutions.
In this way it is possible in bokmål to distinguish between a generic and specific reading: den hvite mann with an indefinite noun means the Caucasian race in general, while den hvite mannen would refer to a specific white man. Similarly, det hvite huset could be any house of that color.
Two nouns that are coordinated in a more or less conventional fashion may have the bare form without an article.
Note that this does not apply only to formal coordination, but also to other conventionalized combinations of nouns, as in (14D). The definite form is also possible in these cases. Without the coordination or sequence, the definite form is obligatory.
In the standard languages proper names do not take a definite suffix, but they may be modified by an adjective in the definite form, and thus take the prenominal definite article.
With inalienable possession, the possessive relation can be expressed by the definite form of the noun alone.
When an adjective modifies a CP, presumably adjoined to the CP, it is preceded by the definite article.8
Most Norwegian dialects (except in the South East), and some Northern Swedish dialects have a preproprial determiner which is used whenever a proper name (first name) is used as an argument. This determiner typically has the same form as the unstressed pronoun.9
The preproprial determiner is also used with terms for close relatives, such as father and mother, grandparents, and a few others.
Nominal arguments are DPs, and the fact that the preproprial determiners are used only when the name is an argument, indicates that they are merged in D. Preproprial determiners are not used in the standard languages.
In the Danish dialect of West Jutlandic there is no movement of N to D, which means that the definite article is always preposed, whether the noun is preceded by a modifier or not.
This is the only Scandinavian variety without a suffixed definite article.
Demonstratives are the distal den ‘that’ and the proximal DN denne S denna, den här ‘this’. They agree with the head noun for gender and number. The distal den has the same form in writing as the preposed definite article, but it is pronounced with a full vowel, whereas the definite article is pronounced with a reduced vowel. In Swedish, denna is quite formal, and is used with the noun in the indefinite form.
Modifiers occur between the demonstrative and the noun, and adjectives have the definite form in -e or -a.
The demonstrative is merged in D, and thus blocks the movement of the noun to D. N stays in I, and in Norwegian and Swedish it usually has the definite form, creating again double definiteness. In Danish, as we have seen, there is no definiteness feature in I, and the noun therefore stays indefinite.
Like modifiers, demonstratives can also be used without the nominal head, even with complements.
The neuter form det ‘that’ of the distal demonstrative can take a CP directly as its complement.
In colloquial speech and also in informal writing the demonstratives may be reinforced by DN her S här ‘here’ and DN der S där ‘there’, either immediately following the demonstrative, or at the end of the phrase.
In Swedish den här is conventionalized as the proximal demonstrative; see (22).
This use of the prenominal pronouns creates a psychological distance, in the sense that the speaker does not personally know or like the person referred to. Syntactically there seems to be a difference between Danish and Swedish on the one hand, and Norwegian on the other. In the former two languages, the pronoun may precede another determiner, and must be generated above DP. In Norwegian it occupies the determiner position D. In (27D) the pronoun is followed by a definite noun, which in Danish is possible only when the noun is in D, and in (27S), a definite article follows the pronoun. This is much less felicitous in Norwegian:
The case of the pronoun in these constructions seems to be independent of the syntactic function of the phrase. In Danish it is regularly in the accusative, whereas in (p.30) Norwegian and Swedish, the nominative is mostly used, whether the phrase is a subject or a complement.
It seems as if one of the case forms of the pronoun has been coopted as a special strong form for use in these functions.
Interrogative phrases are introduced by an interrogative word (wh-word) in D. The common Scandinavian interrogative determiner is DNb hvilken S vilken ‘which’.
This word is not used in standard nynorsk. Instead of hvilken/vilken, the interrogative pronoun D hvad Nb hva Nn kva S vad ‘what’ can be used in combination with the preposition DN for S för ‘for’ followed by the questioned DP in the indefinite form. In nynorsk this is the regular form.
In nynorsk, kva can be used by itself, or the for-phrase may be left in situ while kva is fronted.
The possessor is generated in SpecNP. It may have the form of a possessive determiner, a genitive phrase, or a prepositional phrase. A possessed noun phrase is necessarily definite, but the definite marking in the phrase depends on the type of possessive construction.
The Scandinavian languages and dialects exhibit a great diversity of possessive constructions, with postnominal or prenominal possessors. All the varieties have prenominal possessors, but only Norwegian among the standard languages (along with certain Swedish dialects) has postnominal possessors as well.
22.214.171.124 Possessive determiner
Possessive determiners for 1st and 2nd person singular and for 3rd person reflexive are inflected for number and gender in agreement with the head noun; 3rd person non-reflexive determiners are not inflected; and for 1st and 2nd person plural it depends on the language or dialect. As a general rule, the forms ending in –s (and the nynorsk forms ending in –ar or –ra) are uninflected. The others agree with the possessed noun for gender and number. The common gender or masculine forms for each possessor are as follows:
In these constructions the possessive determiner remains in SpecNP, while the head noun moves to I to join with the definite suffix, and then on to D.
If there is a modifier, the N-movement is again blocked, as described above in 2.3.1. The definite noun stays in I, and the definiteness feature is lexicalized by the independent definite article in D.
The possessive determiner may also move from SpecNP, via SpecIP to SpecDP, and thereby end up first in the phrase. This is obligatory in Danish and Swedish, which do not have postnominal possessors. This movement is not blocked by an intervening modifier.
The possessed noun never gets the suffixed definite article, but adjectives have the definite form, and the possessed phrase has definite reference.
Possessive determiners may be prenominal in Norwegian under specific circumstances. Postnominal possessives require the definite article on the noun, so if the noun is a proper name, or if it is not expressed, the possessor is prenominal.
The only exception to this generalization are kinship terms, see section 126.96.36.199.
A possessive determiner with contrastive focus may be pre- or postnominal in Norwegian.
The postnominal possessive determiner has properties similar to those of weak pronouns. This means that it cannot project to an extended phrase (Lødrup 2011). So when the pronoun is coordinated, (40), or modified, (41), it is obligatorily prenominal even in Norwegian.
188.8.131.52 Genitive phrase
Genitive phrases are DPs with the phrasal clitic –s added to the last word of the phrase. They are always prenominal in the standard languages. The –s is generated in D, and is cliticized to the possessor phrase which moves from SpecNP to SpecDP. The genitive suffix does not exist in most Norwegian dialects, and is prescriptively avoided in nynorsk (except with proper names).
Again, the possessed noun never gets the suffixed definite article, but adjectives have the definite form, and the possessed DP has definite reference.
184.108.40.206 Possessor doubling construction
Those varieties of Norwegian that tend to avoid the genitive phrase with the clitic –s, have an alternative construction available, the ‘possessor doubling construction’.11
The construction consists of a preposed possessor with the definite suffix followed by the reflexive possessor sin, which is inflected for gender and number in agreement with the possessed noun. This word has no reflexive meaning in this construction. Synchronically, it can be seen as a variant of the gentive –s, similarly generated in D. Diachronically, it started out in Bergen (initially based on a borrowing from Low German), and was then adopted into western dialects and nynorsk as a replacement for the lost genitive –s. Nowadays it has spread to other parts of the country, and is common in speech, and in informal writing even in bokmål.
220.127.116.11 Prepositional phrase
Yet another possessor construction is used in Norwegian, a postnominal prepositional phrase with the preposition til or Nn åt ‘to’, followed by the possessor in the form of a DP, including proper names, but not pronouns.
As with the other postnominal possessors, the possessed noun is in the definite form. The prepositional phrase is generated in SpecNP like the other possessors, where it stays, while the noun moves to I to get the definite article. The prepositional possessor is used in both nynorsk and bokmål as synonymous variants of the genitive (p.36) phrase, or the possessor doubling construction. In some Norwegian dialects this is the only option with a DP possessor.
Postnominal PPs with other prepositions, such as på ‘on’ in N taket på huset ‘the roof of the house’ exist in all the languages, and are considered complements of N, see 2.1.2 above.
18.104.22.168 Preproprial genitive
Dialects that use a preproprial determiner with proper names and kinship terms for close relatives (see (19)), have yet another postnominal possessive expression at their disposal: a 3rd person possessive determiner followed by the name or the kinship term.
From the dialects, these constructions have found their way into nynorsk writing.
22.214.171.124 Kinship nouns
Kinship nouns may form possessive constructions differently from those of other common nouns in Norwegian. In bokmål, and in the kind of urban speech which most closely reflects standard bokmål, kinship nouns may have an unstressed, non-focal prenominal possessive determiner.
In most other dialects and in nynorsk, the possessive is postnominal, as with other nouns, but with the difference that nouns for close relations, mostly core family members, lack the definite article. This is also the case with a postnominal possessive prepositional phrase.
The kinship noun can also be used in an argument position in the indefinite form without a possessive. Then the implied possessor is generally the speaker. If it has the definite form, the implied possessor can be anybody else, but necessarily a 3rd person.
These interpretations depend on a normal context perspective.12
126.96.36.199 Interrogative possessor
Asking about the possessor, Danish and Swedish use an interrogative possessive pronoun, D hvis S vems ‘whose’.
In Norwegian this is more problematic. In conservative literary bokmål, the Danish hvis may still be used, but it is becoming quite obsolete. With varying degrees of formality, and depending on the stylistic level, and the region of the country, interrogative versions of the strategies presented in 188.8.131.52–4 above are used. In informal bokmål and in speech influenced by bokmål, the genitive –s is added to the interrogative pronoun hvem ‘who’, which then corresponds to the Swedish form, (51a). Another solution is to use the interrogative in the possessor doubling construction, (51b). Finally, the prepositional possessor can be used with preposition stranding, (51c).
184.108.40.206 Distributive possessive
Scandinavian has a distributive possessive construction consisting of the quantifier DNb hver Nn kvar S var ‘each’ in combination with the possessive determiner.
Distributive possessives are anaphors, typically bound by the subject. The possessive determiner does not, however, need to agree with the antecedent for person. Even if the antecedent is 1st or 2nd person, the 3rd person reflexive sin may be used. This is the rule in Swedish, whereas in Danish and Norwegian, a 1st or 2nd person pronoun may also be used.
In Norwegian, the quantifier too may inflect for gender in agreement with the possessed noun. This seems to be obligatory when there is person agreement in the possessive (Vangsnes 2002). This results in the following three possibilities with a 1st person subject and a possessed neuter noun.
The pronouns constitute a varied class of words. They are the only words which have maintained a case distinction in Scandinavian. Pronouns have syntactically much in common with definite noun phrases. I will assume that they are generated in N, from where they move to I, and then to D, like definite nouns.
2.4.1 Personal pronouns
There are personal pronouns for three persons and two numbers in all the Scandinavian languages. In addition there is a reflexive pronoun for the 3rd person, both singular and plural.
220.127.116.11 Forms and use
There are two cases in the standard varieties, nominative and accusative. In addition there are genitive forms used as possessive determiners (see 18.104.22.168). In some of the varieties there are certain syncretisms between the cases. The spoken varieties distinguish phonologically between a strong and a weak form of the personal pronouns. This distinction is not reflected in the standard orthography in any of the languages, but strong and weak forms may differ syntactically. This will be accounted for in the relevant chapters and sections to follow.
Personal pronouns for the 1st and 2nd person, and the reflexive, are suppletive forms of Proto-Germanic origin.
The nominative is used for the subject of finite verbs, the accusative as complements of lexical verbs and prepositions, and as the subject of non-finite verbs (e.g. in small clauses, see 5.5).
As predicate complements of copula verbs Swedish uses the nominative, which is also the general rule in Norwegian nynorsk, whereas Danish and Norwegian bokmål use the accusative, see 22.214.171.124.
For the 3rd person non-reflexive there are two sets of pronouns. There is a set of pronouns starting in h-, with only singular forms, which is used with human antecedents. For non-human antecedents, and for all plurals, the forms are recruited from the determiners.13 Standard nynorsk, and most Norwegian and some Swedish dialects, use the h-forms for non-human antecedents as well.
The 3rd person pronouns can be presented as follows:
In Danish the subject is in the accusative when it does not refer anaphorically or deictically to an antecedent in the context or situation, but is identified by an apposition or by a relative clause.
In colloquial Danish and Norwegian, coordinated pronouns (especially 1st person) may occur in the accusative case even as subjects.
This is of course considered substandard and frowned upon by prescriptivists.
126.96.36.199 Uses of the neuter singular
The neuter pronoun det has various uses besides referring to neuter nouns. It is used as a non-referential subject with impersonal verbs.
Constructions with non-referring det as a subject will be treated in detail in 6.2.3.
The neuter pronoun is also used to refer to a non-specific nominal of any gender or number (hund ‘dog’ is common gender in Danish, kniv ’knife’ is masculine in Norwegian; see also 7.2.9 on topic doubling).
The neuter pronoun det is used to refer to a sentence or a clause.
Finally, det is the pronoun used when the antecedent is a VP. If the head of VP is a copula or an auxiliary, or ha(ve) ‘have’, det is the complement of this verb. If the head is a lexical verb, it is replaced by the pro-verb D gøre Nb gjøre Nn gjere S göra ‘do’.
The pronoun det necessarily pronominalizes the entire VP, except possible free adjuncts.
The pronominalization in (10b) is impossible, since heim ‘home’ is a complement of the verb. However, (10d) is fine, since i går ‘yesterday’ is not a complement, but an adjunct, and thus not included in the antecedent of det.
188.8.131.52 The reflexive pronoun
Scandinavian has a reflexive pronoun for 3rd person: DS sig N seg. This is used both for the singular and the plural. It is an anaphor, and has no nominative form.
For 1st and 2nd person reflexives the regular personal pronouns are used.
The reflexive pronoun is also part of several lexicalized verbal expressions, where it has no semantic argument role.
The word DNb selv Nn sjølv S själv ‘self’ may be combined with the reflexive or the personal pronoun. This is used primarily when the reflexive is an argument on a par with more natural non-reflexive arguments, and when it is locally bound, that is, an argument of the same predicate as the antecedent.
2.4.2 Interrogative pronouns
Interrogative pronouns are DNb hvem Nn kven S vem ‘who’ and D hvad Nb hva Nn kva S vad ‘what’.
2.4.3 Indefinite pronouns
The indefinite pronouns are man and DNbS en Nn ein ‘one’. The former is nominative only, and can only be used as a subject of finite verbs. For the accusative en/ein ‘one’ is used. In Norwegian en/ein is also used as a subject. There is also a genitive form in –s.
2.4.4 Pronominal phrases
Pronouns of 1st and 2nd person plural may take a complement in the form of a plural NP.
A pronoun of any person may take a prepositional phrase as a complement.
In Norwegian and Swedish, the nominative form of a pronoun with a complement is frequently used even in positions where the accusative would be expected.
In very restricted contexts a personal pronoun may be modified by an adjective. This usually has an emotional effect, and is mostly used in exclamations or in addressing people in an endearing way.
Since personal pronouns are definite, these adjectives have the definite form (stakkars ‘poor’ is uninflected). The pronoun may be in the nominative (preferred in Swedish), or in the accusative (preferred in Danish).
2.5 Relative clauses
Relative clauses generally follow all other elements of the DP, although there is a certain freedom of order among elements following the noun. Relative clauses are introduced by the complementizer som, which may be deleted, or by a relative pronoun. On the analysis adopted here, the relative clause is right-adjoined to its correlate. The internal structure of relative clauses is discussed in section 8.3. Restrictive and non-restricitve relative clauses are syntactically distinct.
2.5.1 Restrictive relative clauses
A restrictive relative clause is right-adjoined to IP, which then is its correlate.
The relative clause is adjoined to the lowest IP node, below any later adjoined adjective. This can be seen by the fact that a polarity item such as ‘ever’ inside the relative clause may be licensed by an adjective in the superlative (Platzack 2000, Julien 2005).
This is possible only if the adjective c-commands the relative clause. The D-projection is consequently not part of the correlate of restrictive relative clauses.
Restrictive relative clauses are really modifiers, and may therefore be combined with the preposed definite article, even if there is no preposed modifier.
In Norwegian and Swedish there is then double definiteness, as with preposed modifiers, while in Danish the noun is without the definite suffix, as shown in 2.3.1. The preposed article may also be missing. In Danish, the noun then has the definite suffix, but the preferred reading is non-restrictive.
An indefinite noun preceded by the definite article in Norwegian or Swedish usually has a non-specific reading.
The nominal correlate may be missing, so that the relative clause is adjoined directly to the determiner, which in such cases normally has a full vowel or stress, and should therefore be considered a demonstrative rather than an article.
A pronoun may also be the correlate of a relative clause.
Note the contrast between the Norwegian sentences in (6) and (7). In (7), the correlate is a plural pronoun in the accusative, while in (6) it is a plural demonstrative, which has no case inflection. The matter is further complicated, however, since in Norwegian a pronoun followed by a relative clause may have the nominative form even in positions where the accusative would be expected, as with prepositional complements (see 2.4.4 (19)).
2.5.2 Non-restrictive relative clauses
A non-restrictive relative clause is right-adjoined to DP, so the correlate is the full DP, including any definite article. The consequence is that a preposed definite article is possible only if there is a modifier in the DP, which blocks N-movement. In other cases, the correlate is a definite noun with a suffixed article. A preposed determiner will have to be interpreted as a demonstrative.
In this case, the adjective does not c-command the polarity item.
The different structural positions of the two types of relative clauses is seen clearly with proper name correlates. A proper name does not normally take a prenominal definite article, so when it makes up a total determiner phrase, (11a), a non-restrictive relative clause does not alter that fact. However, with a restrictive relative clause adjoined to IP, the definite article is required, (11b), as with adjectival modifiers, (11c).
2.5.3 Infinitival relatives
A third type of relative clause are infinitival relatives (described in 8.3.6). These are infinitival clauses with the infinitive marker and a transitive verb (or a verb + preposition).
The complement of the verb (or the preposition) is relativized, and the matrix verb usually has the meaning of possession or acquisition, or their negation.
In a different kind of infinitival relative construction, the relativized complement appears as the subject of the matrix sentence.
This type of infinitival relative is much more common as a complement of adjectives, see 3.2.3.
Determiners, pronouns or definite nouns in D may be preceded by other elements, left-adjoined to DP. There are basically three categories which are found in this position: universal quantifiers, focusing elements, or proximal demonstratives.
2.6.1 Universal quantifiers
The quantifiers D al ‘all’ NS all, DN begge S båda ‘both’, and DNb hele Nn heile S hela ‘whole’ precede a definite noun, the prenominal definite article, or a pronoun.
Note that even in Danish the noun has the definite form, which means that it has moved to D, and that the quantifier is above D, in a higher quantifier phrase. Consequently, possessives and demonstratives can also be preceded by universal quantifiers.
A universal quantifier may be followed by an indefinite DP, but still have definite reference. As a consequence, it cannot occur in existential sentences. Existential quantifiers, such as ‘many’, do not have that effect.
2.6.2 Focusing elements
The word DNb selve Nn sjølve S själva with the basic meaning ‘him-/herself’, sometimes even with a superlative ending, may precede the DP.
Other focusing elements are words and expressions meaning ‘only’, ‘also’, ‘even’, etc.
These are general focusing elements that are used with most phrasal categories.
The proximal demonstrative DN denne S denna ‘this’ may occur in front of the definite article or a possessive determiner.
It seems clear that this demonstrative also has a focusing function.These constructions are extremely rare and of dubious acceptability, especially in Norwegian. They are limited to an archaic style.
2.7 Further reading
Comprehensive descriptions of nominals: Delsing 1993; Julien 2005; Halmøy 2016.
Complement clauses: Lødrup 2004.
The status of the definite article: Faarlund 2009.
Possessives: Vangsnes 2002; Lødrup 2009a, 2011, 2012, 2014a.
Pronouns: Josefsson 2009.
Pronouns used as prenominal determiners: Johannessen 2008; Johannessen and Garbacz 2014.
The use of ‘self’: Hellan 1988 (chapter 2); Lødrup 1999, 2007.
Relative clauses: Platzack 2000.
(1) For a discussion of the morphological status of the sufixed definite article, see Faarlund (2009) and Wetterlin and Lahiri (2015). In the following I will use the term suffix for the postnominal definite article.
(2) I collapse the two projections Num and n of Julien (2005) into one. The nominal head I then corresponds to T in the extension of VP (see chapter 6), which also contains more than one feature category, namely tense and mood.
(3) This is not the only solution. Delsing (1993) treats the adjective as a head of an AP that takes the NP as its complement, but treating the adjective as a head raises empirical problems. Julien (2005) has the adjective phrase as a specifier in a designated α-projection. Structurally, this amounts to the same as my adjunction.
(4) Common gender indefinite singular has a null suffix throughout, and will not be marked in the glosses. Even in varieties with three genders, there is generally no masculine/feminine distinction in adjectives.
(6) The question of why this movement is blocked is still unresolved. Various proposals depend on whether the adjective is a phrase or a head, and whether the noun movement is head or phrasal movement. With the adjective treated as a head, head movement of the noun would be blocked by the Head Movement Constraint (Delsing 1993, Sandström and Holmberg 1994, Vangsnes 1999). But as we have seen, the modifier is a phrase, and can therefore not block head movement. Julien (2005: 29f.) argues that the movement of the noun is phrasal movement, which is possible because the adjective in her analysis is not in the direct projection line above IP, but has its own projection. Her proposal implies that the noun ends up in SpecDP, and not in the D head, where the definiteness feature is located. For a discussion within the framework of distributive morphology, see Hankamer and Mikkelsen (2005). I will not discuss these theoretical problems and their possible solutions any further.
(7) The definite declensional ending in adjectives will generally not be marked in the glosses.
(13) These pronouns are still identical to the determiners in their written forms, except the plural accusative dem, and so with non-human reference it may be difficult to decide whether they should be analyzed as pronouns or as determiners.