Phrasal Proper Names in German and Norwegian

Abstract

This paper discusses the morpho-syntax of phrasal proper names like Deutsche Bahn ‘German Railway’ and Norske Skog ‘Norwegian Forest’ in German and Norwegian. As regards determiner elements, there are three types of phrasal proper names in German: some proper names do not have a definite article, some do, and yet others exhibit a possessive. Depending on the syntactic context, the first two types pattern the same as regards the presence or absence of the article but contrast with the third, where the possessive is always present. It is proposed that proper names in German vary in their structure as regards the presence of the DP-level: unlike articles, possessives have a referential marker, and a DP is obligatorily projected with the latter element. Norwegian is different. While proper names in Norwegian also vary in the presence or absence of determiners, there is no flexibility—determiners are always present or always absent, independent of the syntactic context. It is proposed that unlike in German, the DP-level in Norwegian is always present. As argued by Roehrs (Glossa J Gen Linguist, 5(1):1–38, 2020, https://doi.org/10.5334/gjgl.1267), phrasal proper names involve a regular syntactic derivation. Given that elements of regular DPs are sensitive to definiteness in Norwegian, it is proposed that Norwegian proper names involve an obligatory definiteness feature. As this feature surfaces in the DP-level, the latter must be present in that language in all instances. Besides this cross-linguistic difference, we document that phrasal PN may show features of recursivity evidenced most clearly in Norwegian.

1 Introduction

Proper names (PN) have been of interest to many scholars. Linguistically, PN such as (1a) and (2a) have received a lot of attention (e.g., Longobardi 1994; Matushansky 2008; Muñoz 2019). We will label this type inherent pn. PN such as (1b) and (2b) have not been discussed much (but see Allerton 1987; Roehrs 2020). We will call the latter phrasal pn:

(1)	a.	Peter			(German)
		Peter
		‘Peter’
	b.	Deutsche	Bahn	[company]
		German	Railway
		‘German	Railway’
(2)	a.	Per			(Norwegian)
		Peter
		‘Peter’
	b.	Norske	Skog	[company]
		Norwegian	Forest
		‘Norwegian	Forest’

Phrasal PN in German and Norwegian are hybrid elements: semantically, they are referential like inherent PN; unlike inherent PN though, they have descriptive meaning similar to ordinary DPs (Lyons 1999, 21).¹

Syntactically, PN in German take the shape of ordinary DPs (3a). Indeed, they can be quite complex (3b), which includes syntactically indefinite patterns (3c):

(3)	a.	Die	Neue	Frau			[magazine]	(Ge)
		The	New	Woman
		‘The New Woman’
	b.	Institut	für	Sprach	und	Literaturwissenschaft	[department]
		Institute	For	Language	And	Literature.science
		‘Institute for the Study of Language and Literature’
	c.	Ein	Himmel	voller	Betten		[store]
		A	Heaven	Full.of	Beds
		‘A Heaven Full of Beds’

Similar facts hold for Norwegian:

(4)	a.	Den	lille	sjokoladefabrikk-en				[company]	(Nw)
		The	Little	Chocolate.factory-def
		‘The Little Chocolate Factory’
	b.	Institutt	for	lingvistiske	og	nordiske	studier	[department]
		Institute	For	Linguistic	And	Nordic	Studies
		‘Institute for Linguistic and Nordic Studies’
	c.	En	Liten	Øl				[company]
		A	Small	Beer
		‘A Small Beer’

We will focus on the syntactically definite PN and only briefly discuss syntactically indefinite PN. Thus far, it appears that German and Norwegian pattern the same.

However, as we will lay out in the course of this paper, there are also a number of striking differences. One of the major distinctions between the two languages involves the obligatory presence of determiners in German vs. in Norwegian when phrasal PN occur in argument position:

(5)	a.	*(Die)	Deutsche	Bahn	ist	eine	große	Eisenbahngesellschaft.		(German)
		the	German	Railway	is	a	big	railway.company
		‘German Railway is a big railway company.’
	b.	Norske	Skog	er	en	stor	produsent	av	publikasjonspapir .	(Norwegian)
		Norwegian	Forest	is	a	big	producer	of	publication.paper
		‘Norwegian Forest is a big producer of printing paper.’

Note that while the phrasal PN in German seems to behave like a typical singular count noun phrase in that it requires the presence of a determiner, its Norwegian counterpart does not. Conversely, when phrasal PN appear in non-argument position such as part of a complex compound, German does not allow an article but Norwegian does:

(6)	a.	der	junge	“(*Der)	SPIEGEL“-Journalist		[magazine]	(Ge)
		the	young	The	Mirror-journalist
		‘the young Mirror journalist’
	b.	? den	unge	“Den	Røde	Frakk-en“-ekspeditør-en	[store]	(Nw)
		the	young	The	Red	Overcoat-def-sales.clerk-def
		‘the young Red Overcoat salesclerk’

Turning to morphology, it is well known that adjectives in German inflect in accordance with their syntactic context. For instance, if a definite article is present, adjectives have a weak inflection; if there is no determiner, adjectives have a strong inflection. This is often referred to as the strong/weak alternation of adjectives (for background discussion, see Harbert 2007, 134–35 and references cited therein). Interestingly, we find the same with phrasal PN:

(7)	a.	das	Deutsch-e	Historisch-e	Museum	[museum]	(Ge)
		the	German-wk	Historical-wk	Museum
		‘the German Historical Museum’
	b.	Deutsch-es	Historisch-es	Museum
		German-st	Historical-st	Museum
		‘German Historical Museum’

As such, phrasal PN in German pattern like ordinary DPs morphologically as well.

Some (other) interesting differences emerge in Norwegian, where weak endings on adjectives are possible, even when a definite determiner is absent. Compare German (8a) to Norwegian (8b). Furthermore, a weak adjective can also be followed by a strong one under certain conditions (8c):

(8)	a.	* Deutsch-e	Historisch-e	Museum		(Ge)
		German-wk Historical-wk Museum
	b.	Stor-e	norsk-e	leksikon	[reference book]	(Nw)
		Big-wk	Norwegian-wk	Encyclopedia
		‘Big Norwegian Encyclopedia’
	c.	Ny-e	Norsk-Ø	Tilhengersenter	[company]
		New-wk	Norwegian-st	Car.trailer.center
		‘New Norwegian Car Trailer Center’

Besides these syntactic and morphological differences between German and Norwegian, we will illustrate other, related distinctions of these types of phrasal PN. All these differences can be summarized as follows. If a possessive determiner is part of the PN in German, it must be present in all contexts. The definite article fares differently in this language: it must be left out in compound-type structures, in vocatives, or when a demonstrative or possessive is present; it can be left out in certain kinds of listings; and it must be present when the PN occurs in argument positions. In Norwegian, there is no such variation—whatever shape the phrasal PN appears in in the original context, the environment showcasing the official name, its form will not vary in other environments. We will propose for German that phrasal PN with a possessive are frozen at the DP-level. In contrast, phrasal PN with or without a definite article are frozen below the DP-level. As for Norwegian, phrasal PN are invariably frozen at the DP-level. We propose that the latter language involves an obligatory definiteness feature.

In the context of this proposal, other differences involving more complex phrasal PN will be discussed. Both languages allow, with certain qualifications, phrasal PN to be embedded under the adjective for ‘new’ and a few others. This yields a new phrasal PN referring to a different entity. Since the newly formed PN are derived on the basis of other PN, we propose that PN formation is recursive. Again, both languages do not pattern the same here. Unlike neu- ‘new’ in German, nye ‘new’ (with a weak ending) in Norwegian can be followed by a possessive or an adjective with a strong ending, leading to some surface strings that are not possible in regular DPs. We suggest that these unexpected strings indicate different domains within the larger nominal structure. These domains vary cross-linguistically, being a function of the vocabulary elements inside the Lexical Entries. The facts in Norwegian will follow from the proposal that phrasal PN in that language are frozen at the DP-level and that nye forms the head of a functional phrase that selects an existing proprial DP. With German not exhibiting any morpho-syntactic indications of different domains, we propose that the latter language involves a regular but complex nominal.

More generally, we will point out that given that proprial articles can and sometimes must be left out with phrasal PN in German (indicating the absence of the DP-level) and that phrasal PN are clearly recursive in Norwegian (where the higher DP-level involves no referential element), it seems unlikely that referentiality originates in the DP-level.

The paper is organized as follows. In Sect. 2, we review some German data that will set the stage for the subsequent comparison with Norwegian. After discussing a recent proposal in Sect. 3, we turn to the discussion of Norwegian in Sect. 4, illustrating certain differences to German. On the basis of these distinctions, Sect. 5 refines the previous proposal. We close the paper in Sect. 6.

2 DP-Level of Phrasal PN in German

Phrasal PN in German exhibit differences in the DP-level. Depending on the syntactic context, we can identify three types of PN in German as regards determiners. In the so-called original context, some PN have no definite article, some do, and yet others have a possessive element. In other non-argument environments, the presence of the definite article varies: the article can be added or left out in list contexts, but it must be absent in compound-type structures and vocatives. We will also see that the definite article can be replaced by a demonstrative or a possessive. Finally, in argument position, determiners are obligatory. After illustrating these points, the different properties of these PN are summarized in Table 1.

Table 1 Presence of determiner elements in phrasal PN in German

Before we start, a brief note on spelling in German is in order (also Nübling et al. 2015, 86ff). With the exception of determiners, prepositions, and particles, the elements of PN are usually capitalized. In fact, sometimes a whole element involves capital letters (e.g., Der SPIEGEL ‘The Mirror’). However, capitalization is not entirely consistent and cannot be used to identify what belongs to a PN (cf. the capitalization of the adjective in (9b) vs. (11b) below). In the gloss, we indicate by capitalization what belongs to the PN. The translation follows the gloss closely. With a few exceptions, this includes capitalization.

2.1 No Determiner in Original Context

We start with the context where complex PN appear in public, that is, on company logos, name signs, etc. Following Roehrs (2020), we will refer to this environment as original context, and we take this string to be the official name. As seen in (9), some PN do not have a determiner:

(9)	a.	stern		[magazine]	(Ge)
		Star
		‘Star’
	b.	Neue	Post	[magazine]
		New	Mail
		‘New Mail’

The absence of the determiner seems to be confirmed by contexts where PN occur as part of compounds (10a) (Schlücker 2018).² Furthermore, it is possible to add a definite determiner such as a definite article or, importantly, a demonstrative (10b). Finally, a possessive determiner can also be added (10c):

(10)	a.	der	junge “(*die)	Neue	Post“-Journalist					(German)
		the	young the	New	Mail-journalist
		‘the young New Mail journalist’
	b.	Mann,	die / diese	Neue	Post	ist	verdammt	teuer	geworden!
		man	the / this	New	Mail	is	darn	expensive	become
		‘Boy, the / this New Mail has become darn expensive!’
	c.	Ach,	meine	Neue	Post	fehlt	mir	aber!
		oh,	my	New	Mail	misses	me	prt
		‘Oh, I am really missing my New Mail!’

2.2 Definite Article in Original Context

There are also some PN that have a definite article in the original context:

(11)	a.	Der	SPIEGEL		[magazine]	(Ge)
		The	Mirror
		‘The Mirror’
	b.	Das	kleine	Weinlokal	[restaurant]
		The	Little	Wine.pub
		‘The Little Wine Pub’

Note though that this article must be left out in compound contexts as well (12a). In fact, it can be replaced by a demonstrative (12b) or a possessive determiner (12c):

(12)	a.	der	junge “(*Der)		SPIEGEL“-Journalist				(German)
		the	young	The	Mirror-journalist
		‘the young Mirror journalist’
	b.	Mann,	dieser	SPIEGEL	ist	verdammt	teuer	geworden!
		Man	this	Mirror	is	darn	expensive	become
		‘Boy, this Mirror has become darn expensive!’
	c.	Ach,	mein	SPIEGEL	fehlt	mir	aber!
		oh,	my	Mirror	misses	me	prt
		‘Oh, I am really missing my Mirror!’

Observe that with the exception of the original context, this type of PN behaves the same as that discussed in the previous subsection. Specifically, a definite article can be absent in non-argument position (12a), but a determiner including a definite article (see Subsect. 2.4) must be present in argument position (12b,c). In view of these different distributions, we point out already here that the presence of the definite article in German varies according to syntactic context (e.g., non-argument position vs. argument position).

2.3 Possessive in Original Context

There is a third type of phrasal PN that exhibits a possessive, either a possessive determiner or a Saxon Genitive:

(13)	a.	Dein	Telefonladen	[store]	(Ge)
		Your	Phone.store
		‘Your Phone Store’
	b.	Conny’s	Container	[store]
		Conny’s	Container
		‘Conny’s Container’

Contexts involving compounds pattern differently here. While perhaps not entirely perfect, the possessive determiner can be present in such contexts (14a). In fact, the possessive determiner cannot be exchanged by a definite article, a demonstrative, or a different possessive (14b) without changing the status of the PN (# indicates the loss of PN status):

(14)	a.	?	der	junge	“Dein	Telefonladen“-Verkäufer			(German)
			the	young	Your	Phone.store-shop.assistant
			‘the young Your Phone Store shop assistant’
	b.	#	der	/	dieser	/	mein	Telefonladen
			the	/	this	/	my	phone.store
			‘the / this / my phone store’

Note that (14a,b) differ from the two cases discussed above: (14a) has a determiner element present, and (14b) does not allow this determiner to be replaced by another determiner including a different possessive. To be clear, possessive determiner elements in phrasal PN are obligatorily present. To take stock thus far, there are three types of phrasal PN as regards determiners.

2.4 Presence or Absence of the Definite Article in Other Contexts

Besides the original context, there are other non-argument environments PN can occur in. For instance, they can be found in various listings, entries in reference books, etc. In these contexts, the definite article can be added when it is absent in the original context (15a), or it can be left out when it is present in the original context (15b) (also Kolde 1995, 404; Nübling et al. 2015, 81). This yields a syntactically optional article in these types of contexts. In contrast, a possessive cannot be left out without changing the status of the PN (15c):

(15)	a.		(die)	Deutsche	Bank	[bank]	(Ge)
			the	German	Bank
			‘the German Bank’
	b.		(Die)	Neue	Frau
			The	New	Woman
			‘The New Woman’
	c.	#	Telefonladen
			phone.store
			‘phone store’

As far as we have been able to establish, the presence or absence of the definite article in (15a,b) is somewhat random and seems to depend on the individual choice made by the author of the list.

Next, consider vocatives. In this context, a definite article is not possible (16a,b). PN with possessives fare better although they are still a bit marked (16c):

(16)

a.

Hey,

(*die)

Neue

Post,

warum

hast

du

die

Preise

schon

wieder

(Ge)

hey

the

New

Mail

why

have

you

the

prices

already

again

erhöht?

increased

‘Hey, New Mail, why have you increased your prices once again?’

b.

Hey,

(*Der)

SPIEGEL,

warum

hast

du

die

Preise

schon

wieder

erhöht?

hey

The

Mirror

why

have

you

the

prices

already

again

increased

‘Hey, Mirror, why have you increased your prices once again?’

c.

?

Hey,

Dein

Telefonladen,

warum

hast

du

die

Preise

schon

wieder

hey

Your

Phone.store

why

have

you

the

prices

already

again

erhöht?

increased

‘Hey, Your Phone Store, why have you increased your prices once again?’

Finally, PN in argument position are different. Here, a determiner element must be present with all types of phrasal PN (also van Langendonck 2007, 18). This is illustrated with a PN that lacks a definite article in the original context:

(17)	a.	*(Die)	Deutsche	Bank	ist	groß.	(German)
		the	German	Bank	is	big
		‘The German Bank is big.’
	b.	für	*(die)	Deutsche	Bank
		for	the	German	Bank
		‘for the German Bank’

The same holds for PN in the plural, which refer to unique sets of elements or events:

(18)	a.	(die)	Schlesische(n)	Kriege	[historical events]		(Ge)
		the	Silesian	Wars
		‘the Silesian Wars’
	b.	# Schlesische	Kriege	waren	grausam.
		Silesian	wars	were	cruel
		‘Silesian wars were cruel.’
	c.	Die	Schlesischen	Kriege	waren	grausam.
		the	Silesian	Wars	were	cruel
		‘The Silesian Wars were cruel.’

More generally, the obligatory presence of determiners when phrasal PN occur in argument position in singular or plural contexts means that phrasal PN project (definite) DPs; that is, they are subject to regular syntactic constraints.

Abstracting away from adjectives contained in the phrasal PN, the above discussion is summarized in Table 1 ((+) = already present in the original context; [+] = replacing article from original context).

With the exception of the original context (second column in Table 1), the first two types of PN pattern the same. Although treating these two cases the same may yield a simpler account for German, it will become clear in Sect. 4 that the original context in Norwegian fixes the shape of the PN in the latter language (i.e., PN in Norwegian are ‘frozen’ in appearance). As such, we assume that the original context indicating the presence or absence of the definite article is also important in German. This is in agreement with Payne and Huddleston (2002, 518), who argue that English periodicals have no articles as part of their official names (our original context) but that these names are used with articles in most contexts.

Since the definite article can be left out or be exchanged by another determiner, Roehrs (2020) proposes that this proprial article is an expletive element similar to der in der Peter ‘(the) Peter’ (see Longobardi 1994; Nübling et al. 2015, 123ff). Below, we will follow Roehrs (2020) in making a distinction between the proprial core and the left periphery of phrasal PN.³ We will propose that if the definite article does not appear in the orginal context, it is not part of the PN; more precisely, it is not in the Lexical Entry. If the definite article is present in the original context, it is part of the Lexical Entry but not its proprial core (i.e., it has no—what we call for now—referential marker). As such, all definite articles are part of the left periphery of PN. Finally, the possessive is part of the Lexical Entry and the proprial core of the PN. Given this, note already here that the proprial core and the left periphery are not structurally fixed in German—they are relative notions that depend on the type of the determiner present (as we will see below, this is different in Norwegian).

2.5 Some Diagnostics of Phrasal PN

As illustrated above, phrasal PN in German resemble ordinary DPs: morphologically, their inflections vary depending on the context (e.g., the presence or absence of an article; for different morphological cases, see below); syntactically, their appearance differs as regards the presence or absence of the definite article (e.g., when in argument vs. non-argument position) (see also, e.g., Payne and Huddleston 2002, 517; Weber 2004). Schlücker (2018, 282) observes that “[s]peakers encounter new names on a daily basis. This means that only a small percentage of all existing names is known beforehand to language users.” Given these points, we need to provide some diagnostics that justify the distinction between phrasal PN and ordinary DPs.

As already mentioned in the introduction, phrasal PN have some intriguing properties. Allerton (1987, 64–69) provided a brief discussion of some properties of phrasal PN in English. In the context of German, Roehrs (2020) made this discussion more formal and added to it. Some of these points are reviewed below. We focus on the comparison of phrasal PN and definite DPs.

Starting with the semantics, phrasal PN are similar to definite DPs in some respects. As already noted, both have descriptive meaning, they are definite in interpretation, and they are referential. The first characteristic, descriptive meaning, was already discussed and illustrated above. As to the definiteness interpretation, it is well known that definite DPs exhibit a Definiteness Effect in existential there-constructions (19a,b) (Milsark 1974). The same goes for inherent PN (19c) and phrasal PN (20):⁴

(19)	a.		There is a man in the garden.
	b.	*	There is the man in the garden.
	c.	*	There is Peter in the garden.
(20)	a.	*	There is King’s College in Cambridge.
	b.	*	There is Penny Lane in Liverpool.
	c.	*	There is Griffy Lake/Lake Monroe in Bloomington.

Continuing with the third property, both phrasal PN and definite DPs are referential in that the DPs in (21a) and (21b) pick out a referent in the world:

(21)	a.	Das	Deutsche	Historische	Museum	ist	in	Berlin.		(German)
		the	German	Historical	Museum	is	in	Berlin
		‘The German Historical Museum is in Berlin.’
	b.	Der	große	braune	Hund	steht	am	Berliner	Rathaus.
		the	big	brown	dog	stands	at.the	Berlin	City.hall
		‘The big brown dog is standing at the City Hall of Berlin.’

However, the conditions that bring about referentiality here are not the same. The phrasal PN in (21a) refers to the same unique entity independent of linguistic context; for instance, the phrasal PN can occur in different situational contexts, and the reference is still the same. In contrast, the reference of the definite DP in (21b) may vary depending on the situational context (Kroeger 2019, 21). Indeed, Lyons (1999, 21–22) makes the distinction between inherently unique definites and contextually unique definites.⁵ In this vein, Kripke (1971) states that proper names are rigid designators; that is, they refer to the same entity in all possible worlds (also Heim and Kratzer 1998, 304; for detailed empirical discussion of inherent PN as regards rigidity and transparent/de re readings, see Longobardi 1994, 637–39).

In his discussion of referentiality, Longobardi (1994) makes a distinction between (inherent) PN and definite DPs. To begin, he argues that referentiality is related to the D-position. He shows that besides PN (and pronouns), common DPs preceded by a definite article can have specific definite interpretations. He proposes that this interpretation is derived in the D-position. Specifically, he argues that raising N to D yields this interpretation and that the proprial article is an expletive element (Longobardi 1994, 648–50, 655). In the appendix of this work, Longobardi suggests that all D-positions are universally generated with the abstract feature [±R(eferential)]. If [+R] is checked, it yields a grammatical derivation involving referentiality. Importantly, only PN (and pronouns) can check this feature, but common definite DPs cannot. Crucially, the D-position, and thus the presence of the DP-layer, is taken to be responsible for referentiality. That the DP-level is related to referentiality appears to be problematic though. This can be seen with certain cases of anaphoric reference, which will also exhibit another difference between PN and common nominals.

It was pointed out above that the definite article must be left out in compound structures (and vocatives) in German. As is clear from anaphoric reference, PN as part of compounds may still refer, but ordinary nouns may not. Specifically, while it is possible to establish coreference in (22a), this is (usually) impossible in (22b) (Haspelmath 2002, 156–60, 2011, 50–51; Schlücker 2018, 290–92; but see also Ward et al. 1991; Alexiadou 2019, who provide instances where the (b)-type of example is possible; LM stands for linking morpheme). Following Ward et al. (1991), we use italics to indicate the anaphoric relation between the antecedent and the pronoun (but remain agnostic here about the nature of this relation):⁶

(22)

a.

Das

ist

die

neue

“SPIEGEL“-Idee.

Er

hat

sie

scharf

verteidigt.

(German)

this

is

the

new

Mirror-idea

it

has

her

strongly

defended

‘This is the new Mirror idea. It defended it strongly.’

b.

* Das

ist

die

neue

Hund-e-hütte.

Er

hat

sie

nicht

gemocht.

this

is

the

new

dog-lm-hut

he

has

her

not

liked

‘This is the new kennel. He did not like it.’

In other words, phrasal PN do not depend on the presence of the definite article to be referential. As already mentioned, Longobardi (1994) proposes for inherent PN that this element is an expletive article. Above, we sided with Longobardi and assumed that this is also true of phrasal PN. Now, if the obligatory absence of the article coincides with the absence of the DP-level in the compound modifier, then the referentiality of PN is not related to that phrasal layer (but the referentiality of common DPs may well be).⁷

To be clear then, while other elements (e.g., definite DPs) can also be referential, the referentiality of phrasal PN is different. For current purposes, we assume it is inherent with PN, and we will label this semantic property as ‘proprial’ referentiality. Given the above discussion of definiteness, we follow the tradition in the literature (e.g., Payne and Huddleston 2002, 517, 520) and take phrasal PN to be definite in interpretation. Thus, we assume that proprial referentiality entails definiteness.

Turning to the syntactic diagnostics, phrasal PN are frozen. They are fixed as regards morphological number; for instance, if singular, they cannot be pluralized.⁸ Furthermore, restrictive adjectives (but see also below), degree elements, or numerals cannot be added, and elements that are part of the PN cannot be substituted by other elements (e.g., question words). In addition, elements cannot be reordered within the PN, there is no subextraction out of PN, and no elements can be omitted (although the head noun of the PN can be elided).⁹ All these properties have been discussed in detail in Roehrs (2020). Here, we illustrate the frozenness of phrasal PN by the impossibility of adding a degree word (23a) and the impossibility of DP-internal reordering (23b,b’):

(23)	a.	das	(#sehr)	Alte	Testament		[bible]			(Ge)
		the	very	Old	Testament
		‘the Old Testament’
	b.	Conny’s	Containter
		Conny’s	Container
		‘Conny’s Container’
	b.’	# der	Container	Conny’s	/	von	Conny	/	der	Conny
		the	container	Conny’s	/	of	Conny	/	the.gen	Conny
		‘the container of Conny’s’

As to the morphology, we illustrated in the introduction that the strong/weak alternation of adjective endings works the same in phrasal PN as in ordinary DPs. This includes inflections inside phrasal PN (cf. (7a,b)). Furthermore, the proprial article also changes its form according to the syntactic context. Compare (24a) to (24b), where der has been changed to des. Indeed, this may also involve the varying inflection on a noun inside a phrasal PN (24c). Thus, these properties do not differentiate the two types of noun phrases. As far as we know, there is only one (minor) morphological difference between phrasal PN and ordinary DPs that helps distinguish them: ordinary nouns must have a genitive ending (24a) whereas nouns as part of PN may optionally not (24b) (Nübling 2012):¹⁰

(24)	a.	das	Zerbrechen	des	Spiegel-s		(German)
		the	shattering	the.gen	mirror-gen
		‘the shattering of the mirror’
	b.	die	Berichterstattung	des	SPIEGEL(-s)
		the	reporting	The.gen	Mirror-gen
		‘the reporting of the Mirror
	c.	Jahrestagung	des	Institut-s	für	Deutsche	Sprache
		annual.convention	the.gen	Institute-gen	For	German	Language
		‘annual convention of the Institute for the German Language’

This inflectional difference between (24a) and (24b) is presumably due to the presence of a referential marker on the PN in (24b) (see next section). This tendency toward morphological invariability of PN is often referred to as Schemakonstanz (‘schema constancy’) in the German literature (e.g., Nübling 2017).

Finally, beside the above properties, PN can also be recognized in certain semantic/pragmatic contexts; for instance, it is clear that Conny’s Container in (25a) is not some kind of bin but rather a store. Similarly, Der SPIEGEL in (25b) is not an object in the house but rather a magazine:

(25)	a.	Ich	habe	ein	T-shirt	in	Conny’s	Container	gekauft.	(German)
		I	have	a	t-shirt	in	Conny’s	Container	bought
		‘I bought a t-shirt in Conny’s Container.’
	b.	Sie	hat	den	SPIEGEL	gelesen.
		she	has	The	Mirror	read
		‘She read the Mirror.’

The overall picture that emerges is that while phrasal PN resemble ordinary DPs on the surface, they exhibit some distinguishing properties. Most importantly for current purposes, operations that change the syntactic shape of the proprial string are severely restricted. Roehrs (2020) proposes that phrasal PN are frozen to certain syntactic operations. He proposes that phrasal PN have a derivation similar to that of ordinary DPs. However, further derivational options are constrained by the presence of a referential marker that is part of the Lexical Entries of phrasal PN. This referential marker yields a type of freezing of the PN to certain syntactic operations. In the next section, we will review some of the details of Roehrs’ proposal. This will set up the contrastive discussion of German and Norwegian in the later sections.

3 Previous Proposal

As seen above, phrasal PN have special properties. The following proposal has two parts. In the first subsection, we discuss the acquisition of phrasal PN, which we take to be a one-time memorization procedure yielding complex Lexical Entries. The second subsection addresses the derivations of phrasal PN, which may vary depending on the syntactic context (e.g., argument vs. non-argument position).

3.1 Proprialization and Lexical Entries of Phrasal PN

Roehrs (2020) follows Nübling et al. (2015, 16f) in calling this type of PN formation proprialization. He proposes that the current cases involve a memorization procedure that marks a set of lexical and functional items as being (part of) a proper name. This procedure consists of three steps. First, regular lexical and functional vocabulary elements are taken from the lexicon. Second, these elements receive a referential marker. Finally, Proprialization collects these marked elements in a set and stores this set in the lexicon. To be clear, the formation of a phrasal PN is a one-time procedure bringing about the creation of a Lexical Entry, the result being that the language user has acquired the phrasal PN and has stored it in their lexicon (for more details, see the above-mentioned works).¹¹ Consider the resultant Lexical Entries in more detail.

Discussing PN for place names in Dutch such as Amsterdam in (26a), Köhnlein (2015) proposes a tripartite Lexical Entry consisting of a semantic, a syntactic, and a phonological component. Focusing on the first two parts here, he fleshes out the referential marker by positing a referential pointer, a category label, and the feature [+proper].¹² Specifically, the semantic component involves a referential pointer marked by ↑ and a category label indicated by [+settlement] as in (26b). In the syntax, the elements are marked by the feature [+proper] as in (26b’). The individual components interact with one another indicated by the double arrow below ((26b,b’) is taken from Köhnlein 2015, 201):

(26)	a. [Amster1-dam2]3

Semantically, the postulation of a referential pointer yields the unique reference of a PN, and the category label specifies whether this is a name for a person, a city, a company, etc. Syntactically, the feature [+proper] distinguishes these elements from those of an ordinary DP accounting for certain morpho-syntactic differences between the two.

Discussing differences between compound-type PN in Dutch and phrasal PN in German, Roehrs (2020) extends Köhnlein’s (2015) analysis by proposing that phrasal PN involve a list of unordered, multi-component vocabulary elements. This yields a complex Lexical Entry, a fixed set of vocabulary elements stored in the lexicon. As in Köhnlein, these Lexical Entries have three components: a semantic, a syntactic, and a phonological one.

Let us discuss the three types of phrasal PN from Sect. 2 by starting with (27a). To avoid redundancy, Roehrs assumes that there is only one referential pointer present, and he assigns it to the noun along with the category label (27b). The possessive pronominal is a determiner (27c), and the number specification is a feature taken to be in Num, the head of a NumP (27d). The elements marked by [+proper] form the proprial core, that is, obligatorily present elements, of the PN. Stored sets are indicated by curly brackets. Leaving out the category label, a convenient shorthand for (27b–d) is (27e) ([+p] stands for [+proper]):¹³

(27)	a. Dein	Telefonladen	(German)
	Your	Phone.store
	‘Your Phone Store’

The other two types of PN are similar, the main difference being the analysis of the definite article.

Recall that the definite article in Der SPIEGEL ‘The Mirror’ is in the original context (i.e., it is part of the official name) but that it can be left out or be exchanged by another determiner element. Therefore, Roehrs (2020) takes the article to be in the stored set, but it is an expletive as already stated above. As such, the article is not part of the semantic but only of the syntactic and phonological components of the Lexical Entry. Since the article is not obligatorily present in all contexts, he assumes that it does not involve the feature [+proper]. For concreteness, we follow Leu (2015), Roehrs (2013), and references cited therein in taking the article to be a bipartite form consisting of a stem (d-) and an inflection (-er), the latter being determined during the derivation:

(28)	a. Der	Spiegel:	{↑spiegel[+p], d-; [-pl][+p]}
	The	Mirror

The third case, phrasal PN with no article in the original context, is similar to the two previous types. However, here the definite article is absent in all components of the Lexical Entry:

(29)	a. Neue	Post:	{↑post[+p], neu[+p]; [-pl][+p]}
	New	Mail

To be clear, Lexical Entries involve fixed sets of multi-component vocabulary elements. The main difference between the entries above is the presence of the determiner element in all three components (possessive), its presence in the syntax and phonology only (definite article is present in the original context), and its absence in all three components (definite article is not present in the original context).

3.2 Derivation of Phrasal PN

After PN are acquired and stored in the lexicon, they can be used. Depending on the syntactic context they appear in, the derivation of the PN may slightly differ. While the derivation of PN with possessives is the same in all contexts, the analyses of PN with or without a definite article vary according to syntactic context. Specifically, although the derivation is based on the vocabulary elements inside the Lexical Entry, certain minor adjustments are made in specific syntactic contexts.

Roehrs (2020) sides with Allerton (1987) in that phrasal PN involve regular nominal strings on the surface only. As discussed in Subsect. 2.5, these strings do not undergo certain syntactic operations (e.g., DP-internal reordering). In other words, phrasal PN are frozen to certain syntactic processes. Roehrs proposes that PN involve derivations like those of ordinary DPs, but with a few additional contraints. Following Chomsky’s (1995) discussion of non-proprial constructions, he assumes that the elements to be used during the derivation are taken from the lexicon and collected in the Numeration. These elements are then Selected from the Numeration and undergo Merge one element at a time. The derivation proceeds bottom-up. Nouns project NPs, and number features project NumPs. Following Cinque (2005, 2010), adjectives are Merged in Spec,AgrP, and determiners surface in the DP-level (for detailed background discussion of the derivation of DPs, see Julien 2005; Alexiadou et al. 2007, and many others).

Phrasal PN are special in two ways: first, the Lexical Entry of the PN serves as (part of) the Numeration; that is, the individual vocabulary elements of the Lexical Entry become (part of) the Numeration. Second, recall that with the exception of the definite article, every item in the Lexical Entry is marked by the feature [+proper]. Roehrs proposes that during the derivation this feature spreads to the hosting phrase of this item. Specifically, this feature projects from the head to its phrase (30a). Norris (2014, 136f) points out that this type of percolation immediately follows from Bare Phrase Structure (Chomsky 1995). Furthermore, with adjectives located in Spec,AgrP, this feature spreads by Spec-head agreement from the specifier to the head of the hosting phrase (30b) (for the discussion of Spec-head agreement, see Chomsky 1986; Gallmann 1996; Alexiadou 2005; Koopman 2006; Baker 2008).¹⁴ Note that (30a) will yield the fact that the entire phrase containing the specifier will also have that feature:

(30)	a.	X[+p]	->	XP[+p]
	b.	Spec,YP[+p]	->	Y[+p]	(->	YP[+p])

As a consequence, all phrases whose specifier or head have the feature [+proper] will have that feature on the hosting maximal projection as well. This will become relevant below.¹⁵ Other than that, the derivation proceeds in the way briefly outlined above and illustrated in more detail below. Consider the tree diagram for (31a) in (31b):

(31)	a. Dein	Telefonladen:	{↑telefonladen[+p], dein[+p]; [-pl][+p]}
	Your	Phone.store

Given a regular structure, we assume that inflections on nouns, adjectives, and determiners are determined during the derivation in the same way as with ordinary DPs. We take it that the minor differences such as the possible absence of the genitive inflection on the proprial head noun (Subsect. 2.5) are due to the presence of the feature [+proper] on that noun.

The derivations of PN with or without a definite article are slightly different. Starting with PN where the article is part of the original context, we proposed in the previous subsection that the article is part of the Lexical Entry but not the proprial core; that is, it is not marked by [+proper]. This is the case with (32a) derived as in (32b):¹⁶

(32)	a. Der	Spiegel: {↑spiegel[+p], d-; [-pl][+p]}
	The	Mirror

In the context of vocatives, compound-like elements, demonstratives, and possessives, the article must be deleted (in lists, it may be deleted). Note again that this is an expletive element, and its deletion will not lead to a loss of meaning.¹⁷

Third, for phrasal PN where the definite article is not in the orginal context, we followed Roehrs in that the article is not part of the Lexical Entry at all. This involves cases like (33a), which are derived as in (33b):

(33)	a. Neue	Post: {↑post[+p], neu[+p]; [-pl][+p]}
	New	Mail

Unlike above, here a definite article (or other determiners) can be added in certain syntactic contexts. We assume that the relevant determiner is added to the Numeration before the derivation. Specifically, a definite article can be added for list contexts, and a definite article, a demonstrative, or possessive must be added for argument contexts (more on Numerations in later sections). For the latter scenario, we follow Longobardi (1994, 620), who argues that the DP-level is projected for independent reasons when a nominal occurs in argument position (also Stowell 1989). The presence of one of these determiner elements will bring about the DP-level. Note that an added article will be of a definite form because the PN as a whole is referential and thus definite. Consider some consequences of the proposal.

As Longobardi (1994) has demonstrated for Italian, (inherent) PN can undergo certain syntactic operations (e.g., proprial nouns may undergo movement to a location higher than common nouns). Roehrs (2020) proposed that inherent PN are marked by [+proper] as well. This relates inherent PN to phrasal PN directly. This in turn means that the syntactic behavior of phrasal PN can also be taken to be sensitive to the presence of the feature [+proper]. Given that all elements in the syntactic tree are marked by this feature as discussed above, all participate in these syntactic operations equally; that is, no individual element can be singled out. As no element can be affected individually, this yields a type of “freezing” of phrasal PN to certain syntactic operations (Subsect. 2.5). Furthermore, if the feature [+proper] is indeed responsible for this syntactic effect, then it must be present early in the derivation. If it is a marking on a vocabulary element, then vocabulary elements must be inserted early in the derivation (i.e., during syntax), at least with phrasal PN.¹⁸

We have seen above that the definite article can, and sometimes must, be left out. Note that in either case, the PN can still be referential. As discussed in Subsect. 2.5, it is possible for the hearer to pick out the relevant entity in the world. To repeat from above, if the presence or absence of the definite article correlates with the presence or absence of the DP-level, then the referentiality of phrasal PN does not originate in the DP-level (pace Longobardi 1994). If so, referentiality must come from somewhere else, presumably an element in the lower part of the structure. Roehrs follows Köhnlein (2015) in that it comes from the referential pointer on the noun. This concludes the review of a previous proposal. Next, we turn to Norwegian with the intention of refining the proposal.

4 Phrasal PN in Norwegian

To identify some differences between German and Norwegian, we start this section with some basic data of phrasal PN in Norwegian. In the second subsection, we illustrate that unlike in German, phrasal PN in Norwegian are always frozen at the DP-layer. In Subsect. 4.3, we provide our Lexical Entries and the derivations. In the fourth subsection, we show that phrasal PN in Norwegian exhibit characteristics of recursivity that German does not show as clearly. Finally, we propose that the nominal strings of phrasal PN can be delineated in three different domains that vary cross-linguistically in their structural sizes.

Before we present the data, let us provide a brief statement about the spelling conventions of PN in Norwegian. It is generally recommended to spell only the first word of phrasal PNs with an initial capital letter. However, as our examples will illustrate, there is much variation on this point.

4.1 Basic Data

Like German, Norwegian exhibits both syntactically definite and indefinite DP patterns, the difference being that there is more variation with Norwegian PN, just like with regular DPs (for extensive discussion of the latter, see, for instance, Vangsnes 1999; Julien 2005; Schoorlemmer 2012). Starting with the cases involving syntactic definiteness, there are the well-known DP strings that involve a free-standing and a suffixal determiner (34a). In addition, there is also a distributional pattern identical, on the surface, to generic DPs where the suffixal article is missing (34b):¹⁹

(34)	a.	Den	Rød-e	Frakk-en	[store]	(Nw)
		The	Red-wk	Overcoat-def
		‘The Red Overcoat’
	b.	Den	Gaml-e	Major	[restaurant]
		The	Old-wk	Major
		‘The Old Major’

Furthermore, there are phrasal PN that have the patterns involving definite adjectives, sometimes called adjectival determiners. In these cases, the free-standing definite article is missing, but a suffixal determiner may or may not be present. Compare (35a,b):

(35)	a.	Stor-e	Skarv-en	[company]	(Nw)
		Big-wk	Cormorant-def
		‘Big Cormorant’
	b.	Norsk-e	Skog	[company]
		Norwegian-wk	Forest
		‘Norwegian Forest’

Turning to the syntactically indefinite patterns, there is the familiar string involving an indefinite article (36a). In addition, there are phrasal PN that are similar to what have been referred to in the literature on Norwegian as type-denoting DPs (Borthen 2003). The latter lack an indefinite article (36b) (note that the adjective liten in (36a) is a portmanteau form and that Ø in (36b) indicates an assumed null ending):

(36)	a.	En	Liten	Butikk	[store]	(Nw)
		A	Small.st	Shop
		‘A Small Shop’
	b.	Gul-Ø	Sirkel		[company]
		Yellow-st	Circle
		‘Yellow Circle’

Again, all these strings involve well-known patterns. If this is so, then one might expect phrasal PN in Norwegian to contain two adjectives where either both of them exhibit a weak inflection, or both of them show a strong ending. This is indeed the case

(37)	a.	Stor-e	norsk-e	leksikon	[reference book]	(Nw)
		Big-wk	Norwegian-wk	Encyclopedia
		‘Big Norwegian Encyclopedia’
	b.	Fin-Ø	Gammel-Ø	Årgang	[company]
		Fine-st	Old-st	Vintage
		‘Fine Old Vintage’

These are the patterns found in the original contexts in Norwegian. Note that, despite the difference in surface strings, all phrasal PN are referential in the same way; that is, their semantics is independent of their surface shape. With this in mind, we turn to the properties that set phrasal PN in Norwegian apart from those in German.

4.2 Frozen DP-level

In this subsection, we illustrate that phrasal PN in Norwegian can be preceded by demonstratives or possessives (without suppressing the determiner present in the original context). Furthermore, proprial articles in Norwegian are also retained in the context of compound-type structures, vocatives, and list contexts. Finally, Norwegian PN can occur in argument positions without a determiner present. These points are summarized and contrasted with German in Table 2. We will conclude that unlike in German, in Norwegian all phrasal PN are frozen at the DP-level.

Table 2 Presence of determiner elements in phrasal PN in German and Norwegian

Unlike in German, a PN with a free-standing article in Norwegian as in (34a) can be extended by a demonstrative (38a). Similarly, a PN such as in (34b) can take a possessive pronominal on its left (39a). In fact, leaving out the free-standing article of these PN results in degradedness. Consider the two (b)-examples below:

(38)	a.	Wow,	denne	Den	Rød-e	Frakk-en	er	en	flott	butikk!	(Norwegian)
		wow	this	The	Red-wk	Overcoat-def	is	a	great	store
		‘Wow, this Red Overcoat is a great store!’
	b.	?? Wow,	denne	Rød-e	Frakk-en	er	en	flott	butikk!
		wow	this	Red-wk	Overcoat-def	is	a	great	store
		‘Wow, this Red Overcoat is a great store!’

(39)	a.	? Har	dere	besøkt	vår	Den	Gaml-e	Major?	(Norwegian)
		have	you	visited	our	The	Old-wk	Major
		‘Have you visited our Old Major?’
	b.	?? Har	dere	besøkt	vår	Gaml-e	Major?
		have	you	visited	our	Old-wk	Major
		‘Have you visited our Old Major?’

Similarly, adding a demonstrative to (35a) works fine (40a), and the same holds for a possessive pronominal, with (35b) yielding (40b):²⁰

(40)	a.	Så,	dette	Stor-e	Skarv-en	et	er	nytt	selskap.		(Norwegian)
		so	this	Big-wk	Cormorant-def	is	a	new	company
		‘So, this Big Cormorant is a new company.’
	b.	Vet	du	at	mitt	Norsk-e	Skog	har	hjulpet	meg	mye!
		know	you	that	my	Norwegian-wk	Forest	has	helped	me	much
		‘Do you know that my Norwegian Forest has helped me a lot?’

Again, unlike in German, in Norwegian the article can be present in compound-type constructions (41a) and in vocatives (41b):²¹

(41)	a.	? den	ung-e	“Den	Rød-e	Frakk-en“-ekspeditør-en					(Norwegian)
		the	young- wk	The	Red- wk	Overcoat- def -sales.clerk- def
		‘the young Red Overcoat salesclerk’
	b.	Hei,	Den	Rød-e	Frakk-en,	hvorfor	har	du	blitt	så	dyr?
		hey,	The	Red- wk	Overcoat- def	why	have	you	become	so	expensive
		‘Hey, Red Overcoat, why have you become so expensive?’

Similarly, the determiner must also be present in list contexts:

(42)	a.	Den	Rød-e	Frakk-en	(Norwegian)
		The	Red-wk	Overcoat-def
		‘the Red Overcoat’
	b.	Den	Gaml-e	Major
		The	Old-wk	Major
		‘The Old Major’

Finally, note that all these Norwegian PN can, as they are, occur in argument position. Illustrating with the type in (36b), where no determiner of any kind is present, the phrasal PN can be the subject or object of a sentence (43a,b). The same goes for the complement position of a preposition (43c):

(43)	a.	Ny-tt	Image	er	min	favorittbutikk.			[company]	(Nw)
		New-st	Image	is	my	favorite.shop
		‘New Image is my favorite shop.’
	b.	Du	finner	Ny-tt	Image	i	Storgat-a.
		you	find	New-st	Image	in	Big.street-def
		‘You’ll find New Image on Main Street.’
	c.	På	Ny-tt	Image	får	du	hjelp	med	stil-en.
		at	New-st	Image	get	you	help	with	style-def
		‘At New Image you get help with your style.’

Again, this is different in German. Recall that if a definite article is not present in the original context in the latter language, it must be added when the PN occurs in argument position (Subsect. 2.4). With Longobardi (1994), we assume again that syntactic arguments are DPs. We propose that all phrasal PN in Norwegian involve DPs including the surface strings in (35a,b) and (36b), where free-standing determiners are missing.²²

To sum up the discussion of Norwegian thus far, demonstrative and possessive pronominals can be added to the left of PN independent of the presence of determiner elements in those PN. Furthermore, PN with determiner elements can occur in compound-type constructions, in vocatives, and lists. Conversely, no (additional) determiner elements are required when PN are in argument position. For convenience, these points in Norwegian are contrasted schematically with German in Table 2 ((+) = already present in the original context; [+] = replacing article from original context; {+} = determiner present if in original context).

Most importantly, unlike in Norwegian, PN in German may leave out an article in non-argument positions (e.g., compounds, vocatives, lists). Conversely, unlike in German, PN in Norwegian may occur without an article in argument positions. We conclude that given the invariable shape of the PN in Norwegian, the latter are frozen at the DP-level. Next, we consider the stored sets of vocabulary elements and the derivations of phrasal PN in Norwegian.

4.3 Lexical Entries and Derivations

In this subsection, we extend the discussion from German to Norwegian. Most importantly, Lexical Entries are formulated in a similar way, the main difference being the obligatory presence of a definiteness feature marked by [+proper] in Norwegian. Furthermore, derivations are the same, with the obligatory definiteness feature in Norwegian always leading to the projection of the DP-level. Indeed, the marking [+proper] on the definiteness feature results in the freezing of the entire DP to certain syntactic operations.

Starting with the cases involving syntactic definiteness, there are many accounts that seek to explain the distribution of the free-standing and the suffixal determiners as well as the weak endings on adjectives in Norwegian. This is not the place to go into a detailed discussion (but see, e.g., Taraldsen 1990; Embick and Noyer 2001; Julien 2005; Anderssen 2006, 2012, and many others). In order to explain the distribution of free-standing and suffixal determiners as well as weak adjective endings, a number of analyses claim that these elements are definiteness-sensitive and are due to the presence of a definiteness feature or several definiteness features (e.g., Julien 2005; Schoorlemmer 2012; Roehrs 2019). As is well known, German does not have suffixal determiners at all. In addition, the endings on adjectives are not regulated by definiteness either (Harbert 2007, 134–35; Roehrs and Julien 2014). As regards the latter point, consider the following nominal strings, which have the same definite interpretation despite the fact that the article is present in the (a)-examples but absent in the (b)-examples. Importantly, while the adjectives in Norwegian are invariably weak here, they vary in German depending on the presence of the article:

(44)	a.	på	den	best-e	måte-n	(Norwegian)
		in	the	best-wk	way-def
		‘in the best way’
	b.	på	best-e	måte
		in	best-wk	way
		‘in the best way’

(45)	a.	das	folgend-e	Beispiel	(German)
		the	following- wk	example
		‘the following example’
	b.	folgend-es	Beispiel
		following- st	example
		‘the following example’

Thus, unlike in German, the shape of the DP in Norwegian is closely tied to definiteness.

Above, we observed that phrasal PN in Norwegian have different surface patterns: some have determiners, some do not. This means that the frozen DP-level in Norwegian cannot be a function of the presence or absence of determiners. Rather, we propose that what all phrasal PN in Norwegian have in common is the presence of an abstract feature. Given the relevance of definiteness for regular DPs, we take this abstract feature to be a definiteness feature (relating phrasal PN directly to regular DPs). Specifically, we propose for Norwegian that the Lexical Entries for phrasal PN involve a definiteness feature [def] marked by [+proper]. Given the syntactically definite and indefinite strings, we assume that this syntactic feature may have a positive or a negative specification.

One may wonder now why the definiteness feature in Norwegian must be present. Proceeding more tentatively, we pointed out above that, despite the same (referential) semantics, the shape of Norwegian phrasal PN is different. To account for the surface differences, it seems clear that all the relevant vocabulary elements including articles must be present in the Lexical Entries. Notice again that all phrasal PN under discussion here exhibit an overt reflex of (in-)definiteness, articles of various kinds and strong or weak adjectives. To make the discussion concrete, we suggest that all these elements involve an uninterpretable feature for (in-)definiteness. Now, assuming that the aforementioned definiteness feature [±def] involves the interpretable counterpart, its presence in the Lexical Entry and thus in the derivation will allow the uninterpretable feature of the vocabulary elements to be checked resulting in a good derivation.

To sum up, keeping the Lexical Entries as similar to German as possible, there is one substantial difference between the two languages—the presence of the definiteness feature in Norwegian (46a) repeating the example from (34a). Like in German, we take the free-standing article to involve the stem d-; the suffixal article is given here as -e- for concreteness. Note that all the vocabulary elements are marked by [+proper]; that is, they are all part of the proprial core. The Lexical Entry in (46a) can be fleshed out further as in (46b) (u stands for uninterpretable, i indicates interpretable; for Def, see below):

(46)	a. Den	Rød-e	Frakk-en:
	The	Red-wk	Overcoat-def
		{↑frakk[+p], rød[+p], d-[+p], -e-[+p]; [-pl][+p], [+def][+p]}

The Lexical Entry for (34b) is given in shorthand in (47a). The other definite instances are similar: (35a) is like (46a) above, and (35b) is like (47a) below, but in each case the free-standing article is not in the stored set. As for the syntactically indefinite cases (36a,b), we assume a definiteness feature with a negative specification (47b,c):

(47)	a.	Den	Gaml-e	Major:
		The	Old-wk	Major
			{↑major[+p], gammel[+p], d-[+p]; [-pl][+p], [+def][+p]}
	b.	En	Liten	Butikk:
		A	Small.st	Shop
			{↑butikk[+p], liten[+p], e-[+p]; [-pl][+p], [-def][+p]}
	c.	Gul-Ø	Sirkel:
		Yellow-st	Circle
			{↑sirkel[+p], gul[+p]; [-pl][+p], [-def][+p]}

As in German, we claim that the pointer brings about referentiality. (Proprial) referentiality in turn entails semantic definiteness. We assume that the presence of this pointer “overwrites” the negatively specified (syntactic) feature for definiteness in cases like (47b,c). Now, making the standard assumption that the definiteness feature surfaces in the DP-level (Lyons 1999), we can point out that the presence of this feature entails the presence of the DP-level. The feature [+proper] on [def] in the DP-level yields the fact that the DP-level is frozen in Norwegian. With all parts of the tree marked by [+proper], no element can be singled out for certain syntactic operations.

Consider the first steps in the derivation of (34a). Following Taraldsen (1990), Vangsnes (1999), Anderssen (2006, 2012), Julien (2005), and others, we assume that there is a phrase between AgrP and NumP where the definiteness feature [def] originates, call it DefP. With Roehrs (2019), we assume that this feature moves to the DP-level (also Heck et al. 2008, 229; Schoorlemmer 2009):

(48)	a. Den	Rød-e	Frakk-en:
	The	Red-wk	Overcoat-def
		{↑frakk[+p], rød[+p], d-[+p], -e-[+p]; [-pl][+p], [+def][+p]}

Before finalizing the derivation, note again that all the elements of the DP in Norwegian have the feature [+proper]. Consequently, the entire DP is frozen.

Continuing the derivation, the copy of [+def] in DP is spelled out by the free-standing determiner (d-), and the copy in DefP by the suffixal one (-e-). These elements were part of the Lexical Entry and thus formed part of the Numeration. The copy of [+def] in Agr will yield a weak adjective ending (for details, see, for instance, Roehrs 2019). The number feature will be spelled out by the relevant number morpheme (in this case, null). Furthermore, the head noun will raise to Num picking up the number morpheme and will then move on to Def supporting the suffixal article. Finally, regular agreement operations and (late) insertion will bring about the relevant inflections:

As is clear from (48b) and (49), there is some DP-internal movement. This means that the feature [+proper] does not block regular movement operations within the DP (but only displacements due to information structure: focus movement, wh-movement, etc.).²³

More generally, these assumptions explain the difference between German and Norwegian as regards the DP-level. In German, possessives are marked by [+proper] yielding a frozen DP-level; PN with articles, however, are only DPs (necessarily) when they occur in argument position requiring the DP-level to be present for independent reasons; something similar holds for PN without articles. In contrast, Norwegian involves a definiteness feature that is marked by [+proper] thus freezing the DP-level in all cases. With this background in mind, we turn to the facts that are important for the proposal of more complex phrasal PN.

4.4 Recursivity

In Sect. 3, we followed Roehrs (2020) in that phrasal PN involve a regular syntactic derivation. Intriguingly, PN formation is recursive. Norwegian has an element, the adjective nye ‘new’ (and a few others), which can form another PN on the basis of an existing phrasal PN. While German also has this type of element, the properties to be discussed are not as clear there. Besides the differences related to the obligatoriness of the DP-layer, this adds another distinction between the two languages. Note that the following data all lack a determiner in front of the word for ‘new’. In contrast, free-standing determiners are required with regular restrictive modification, a point we will return to at the end of this subsection.

Recall that Norwegian phrasal PN can have two weak adjectives while German cannot (when the definite determiner is absent in the latter language). Consider now (50). What is interesting to observe here is that the Norwegian example in (50a) is ambiguous in interpretation. On the one hand, nye can be part of a regular phrasal PN; in this case, the PN refers to a store that sells new red hats (and presumably some other items). On the other hand, nye can also form a new phrasal PN on the basis of an existing PN (Røde Hatt). In other words, in the first case, a name is given to a store that is established for the first time; in the second scenario, the original store has been remodeled and reopened, or it is possibly under new management in a different location.²⁴ We call the first reading of nye primary interpretation (marked below by the superscript ¹) and the second reading of nye secondary interpretation (indicated below by the superscript ²). As mentioned above, the German pattern is independently out (50b):

(50)	a.	Ny-e	Rød-e	Hatt	[store]	(√Nw1,2)
		New-wk	Red-wk	Hat
		‘New Red Hat’
	b.	* Neu-e	Ägyptisch-e	Museum		(*Ge)
		New-wk	Egyptian-wk	Museum
		‘New Egyptian Museum’

We will see that the different readings correlate with some interesting, unexpected morpho-syntactic facts in Norwegian. Specifically, strings with a secondary interpretation allow the adjective nye (with a weak ending) to be followed by an adjective with a strong ending or by a possessive. This is different from German. To the extent that unexpected patterns are possible in the latter language at all, they are quite marked.

Continuing with the different inflectional options, both languages can have two strong adjectives. Here, the adjective for ‘new’ can only have the primary interpretation in Norwegian. In contrast, German allows both readings:

(51)	a.	Ny-Ø	Gul-Ø	Sirkel	[company]	(√Nw1)
		New-st	Yellow-st	Circle
		‘New Yellow Circle’
	b.	Neu-es	Ägyptisch-es	Museum	[museum]	(√Ge1,2)
		New-st	Egyptian-st	Museum
		‘New Egyptian Museum’

Furthermore, the adjective for ‘new’ with a strong ending cannot be followed by an adjective with a weak ending in Norwegian. This combination seems to be marginally possible for some speakers in German:

(52)	a. *	Ny-Ø	Grei-e	Kafeteria	(*Nw)
		New-st	Nice-wk	Cafeteria
		‘New Nice Cafeteria’
	b.%	Neu-es	Ägyptisch-e	Museum	(%Ge)
		New-st	Egyptian-wk	Museum
		‘New Egyptian Museum’

Conversely, Norwegian nye can be followed by a strong adjective. In this case, nye only has a secondary interpretation. Note also that this surface string is usually ungrammatical in Norwegian, just as it is impossible in German including with PN:

(53)	a.	Ny-e	Grei-Ø	Kafeteria	[cafeteria]	(√Nw2)
		New-wk	Nice-st	Cafeteria
		‘New Nice Cafeteria’
	b.	* Neu-e	Ägyptisch-es	Museum		(*Ge)
		New-wk	Egyptian-st	Museum
		‘New Egyptian Museum’

Before we summarize these four inflectionally different patterns, let us point out that there is a related morpho-syntactic difference between Norwegian and German.

While the weak adjective for ‘new’ can be followed by a possessive in Norwegian yielding a secondary interpretation, this is not possible in German. Again, note that this surface string is usually impossible, also in Norwegian:

(54)	a.	Ny-e	Elses	Blomster	[store]	(√Nw2)
		New-wk	Else’s	Flowers
		‘New Else’s Flowers’
	b.	* Neu-e	Marias	Laden	[store]	(*Ge)
		New-wk	Mary’s	Shop
		‘New Mary’s Shop’

If the adjective for ‘new’ has a strong ending in this context, the resultant nominal in Norwegian is out, but the one in German, while quite marked, is not completely impossible:

(55)	a.	* Ny-tt	Elses	Hjem	(*Nw)
		New-st	Else’s	Home
		‘New Else’s Home’
	b.	?*/?? Neu-er	Marias	Laden	(*Ge)
		New-st	Mary’s	Shop
		‘New Mary’s Shop’

Finally, while the most frequent element in this regard appears to be the word for ‘new’, there seem to be a few other adjectives that combine with existing PN yielding newly formed PN:

(56)	a.	Alt-er	Grün-er	Weg		[street name in Alsfeld]	(Ge)
		Old-st	Green-st	Path
		‘Old Green Path’
	b.	Groß-es	Lexikon	der	Astronomie	[reference book]
		Big-st	Lexicon	the.gen	Astronomy
		‘Comprehensive Lexicon of Astronomy’
	c.	Klein-es	Berlin-Lexikon			[reference book]
		Small-st	Berlin-Lexicon
		‘Little Berlin Lexicon’

Some Norwegian examples are provided below:

(57)	a.	Gaml-e	Oslo	[a borough in Oslo]	(Nw)
		Old-wk	Oslo
		‘Old Oslo’
	b.	Lill-e	Frøen	[a neighborhood in Oslo]
		little-wk	Frøen
		‘Little Frøen’

Note that all these names refer to unique entities that are different from their simpler counterparts, which lack the first element. Thus, these elements are part of the referential component of the PN. More generally, recursivity involving the word for ‘new’ is not an isolated case. However, the elements that can induce recursivity seem to involve a small, restricted set. We continue the discussion with the word for ‘new’.

To sum up thus far, there are some differences between Norwegian and German when the adjective for ‘new’ is followed by another adjective or by a possessive. First, of the four logical inflectional combinations, three are possible in Norwegian with varying interpretative options while German allows only two, one of them being marginally possible for some speakers. Both languages share only one surface string (when both adjectives have a strong ending). Second, when the adjective for ‘new’ has a weak ending, it can be followed by a possessive in Norwegian but not in German. This is summarized in Table 3.

Table 3 Possible inflections on ‘new’ and different following elements

More generally, there are then two unexpected surface strings in Norwegian: the weak adjective nye can be followed (i) by a strong adjective or (ii) by a possessive. These morpho-syntactic patterns correlate with a (necessarily) secondary reading of nye. We take the latter two morpho-syntactic differences to mean that Norwegian nye combines with a lower DP. This fits well with Norwegian phrasal PN being frozen at the DP-level (although the latter is not always overt). Specifically, Norwegian has frozen DPs in isolation as argued in Subsect. 4.2, and we just demonstrated that the very same frozen DPs can occur elsewhere, namely under (secondary) nye yielding recursive patterns. German is different. Assuming that (prenominal) possessives are in the DP-layer, note that German neu- does not combine with a lower DP, but a non-DP is possible; compare (51b), on the secondary interpretation, to (55b).

Before moving on to the next subsection, we would like to point out that the cases discussed above are different from instances involving regular restrictive modification. As pointed out to us by a reviewer, it is indeed possible to distinguish two entities of the same name by adding an intersective adjective such as ‘small’ or ‘big’ (for different types of adjectives, see Kroeger 2019, 275–81). While it is not easy to find authentic examples for phrasal PN, suppose there are two stores or branches with the name Nye Elses Blomster ‘New Else’s Flowers’. In order to differentiate both, one could add a relevant distinguishing adjective, lille ‘small’ in (58a) and store ‘big’ in (58b). Crucially though, a free-standing determiner has to be added to the left periphery as well:

(58)	a.	den	lill-e	Ny-e	Elses	Blomster	(Norwegian)
		the	small-wk	New-wk	Else’s	flowers
		‘the small New Else’s Flowers’
	b.	den	stor-e	Ny-e	Elses	Blomster
		the	big-wk	New-wk	Else’s	flowers
		‘the big New Else’s Flowers’

This makes the current cases different from the above instances, which lack a determiner in front of nye.²⁵ As mentioned above, Nye Elses Blomster refers to a different store than (unmodified) Elses Blomster, and it is not possible to leave out nye with the first name (even if there is only one such store). Indeed, it is not necessary to add an adjective like ‘old’ to the second name to make the relevant distinction (cf. the inherent PN New York vs. York). In fact, it is possible that although both names are linguistically related, they might offer different goods or services. Finally, PN like Nye Elses Blomster typically, but not necessarily, imply the existence of a different store, namely one with a name that lacks the adjective nye. We conclude that the cases discussed in the first part of this subsection do not involve regular restrictive modification (note also in this regard that we have pointed out above that only a handful of adjectives can be used to create new names highlighting the special status of these cases further).²⁶

4.5 Three Domains

In the last subsection, we have shown that phrasal PN are recursive in that Norwegian nye ‘new’ can embed another PN yielding a new one. Interestingly, other elements can appear to the left of that. For instance, the adjective for ‘new’ can be preceded by a non-restrictive adjective and an accompanying determiner (59a). The same holds for German (59b):

(59)	a.	den	berømte	Nye	Elses	Blomster	(√Nw2)
		the	famous	New	Else’s	Flowers
		‘the famous New Else’s Flowers’
	b.	das	berühmte	Neue	Ägyptische	Museum	(√Ge1,2)
		the	famous	New	Egyptian	Museum
		‘the famous New Egyptian Museum’

We propose that the comparative facts from Norwegian, including the varying interpretative options and the unexpected surface strings, follow from the proposal laid out in Sect. 3 and Subsect. 4.3 once some refinements are made. Specifically, the Norwegian data clearly indicate the presence of different domains with phrasal PN. The combinatory options can be summarized as follows:²⁷

(60)		Summary:
	1.	Phrasal proper names have a frozen primary core. Below, this is marked by curly brackets.
	2.	Elements can only be added to the left or right. Focusing on the left side for now, there are two types of additions (leading to the embedding of the primary core):
		a. Creation of a new proper name forming a secondary core (e.g., by adding nye ‘new’ in Norwegian). This is marked by a second set of curly brackets.
		b. Construction of a regular left pheriphery (e.g., by adding a definite article and a non-restrictive adjective: Norwegian det/den berømte ‘the famous’).

Taking the Norwegian example from (59a) above, repeated here as (61a), we illustrate the three domains in (61b). The stored Lexical Entry for a phrasal PN with a secondary core looks as in (61c):

(61)	a.	den	berømte	Nye	Elses	Blomster	(Norwegian)
		the	famous	New	Else’s	Flowers
	b.	γ[ regular left periphery		β{ secondary	α{ primary core }α }β ]γ
	c.	Nye Elses Blomster:
		{↑nye[+p] { blomst[+p], else[+p]; [+pl][+p], [+def][+p]} [+def][+p]}

Linearly, the (regular) left periphery precedes the secondary core, and the latter precedes the primary core. Considering (61a,b), nye is, in a way, a linking element between the two other domains (provided the left periphery is present). Note also that the elements in the left periphery are not part of the stored set in (61c). Similar to German PN that lack a definite article in the original context, we suggest that these items are added to the Numeration along with the vocabulary elements from the Lexical Entry of the PN. To be clear then, the Numeration involves the elements from the Lexical Entry, and non-proprial elements such as non-restrictive adjectives and determiners can be added to it as well. The derivation will proceed on the basis of all these elements (for details, see Sect. 5). It will become clear that the domains are a function of the vocabulary elements in the Lexical Entries. With a definiteness feature present in Norwegian but not German, these domains may vary cross-linguistically in terms of their syntax.

It is worth pointing out that other elements can be in the peripheries. As seen in the discussion of German and Norwegian, unstressed, non-restrictive possessives can also be added in the left periphery (the same holds for demonstratives, not shown here):

(62)	a.	Ach,	mein	Peter	(German)
		oh	my	Peter
		‘Oh, my Peter’
	b.	Å,	min	Per	(Norwegian)
		oh	my	Peter
		‘Oh, my Peter’
	b’.	Å,	Per	min
		oh	Peter	my
		‘Oh, my Peter’

(63)	a.	Ach,	meine	Deutsche	Bank	(German)
		oh	my	German	Bank
		‘Oh, my German Bank’
	b.	Å,	mitt	Norske	Skog	(Norwegian)
		oh	my	Norwegian	Forest
		‘Oh, my Norwegian Forest’

These possessives establish a close emotional relation to the name and its bearer. Finally, non-restrictive relative clauses also appear in the periphery but on the right.

(64)	a.	die	Deutsche	Bank,	die	übrigens	1869	gegründet	wurde
		the	German	Bank	which	incidentally	1869	founded	was
		‘the German Bank, which by the way was founded in 1869’
	b.	Norske	Skog,	som	forresten	ble	grunnlagt	i	1962
		Norwegian	Forest	which	incidentally	was	founded	in	1962
		‘Norwegian Forest, which by the way was founded in 1962’

To take stock, the first point in (60) was derived in Sect. 3 above proposing the operation Proprialization followed by a regular derivation and constrained by the feature [+proper]. The discussion of Norwegian brought to light two differences from German: Phrasal PN in Norwegian are frozen at the DP-level, and they show characteristics of recursivity (not easily seen in German). As discussed in Subsect. 4.3, the first difference followed from the assumption that all stored sets of vocabulary elements in Norwegian involve a definiteness feature marked by [+proper]. The presence of the definiteness feature entails the presence of the DP-level in Norwegian, and the presence of the feature [+proper] explains the fact that the entire DP is frozen. In what follows, we derive the facts involving recursivity and the presence of non-proprial elements in the peripheries by refining the structural assumptions from Sect. 3 and Subsect. 4.3.

5 Domains as Multiple Layers in the Structure

Unlike in German, in Norwegian phrasal PN always involve a frozen DP-level. This will result in similar, but not identical, structures for German and Norwegian. The next two subsections begin by providing some new and some known arguments for the existence of different domains. This will lay the foundation for the postulation of the structures of more complex phrasal PN. The third subsection discusses some extensions. We start the discussion with the structure of Norwegian.

5.1 Structure of Norwegian

Above, we have seen that phrasal PN in Norwegian are frozen at the DP-level. Unlike in German, they can appear in argument position without an article. Furthermore, if an element is added to the left, certain unexpected patterns are possible. For instance, these patterns become visible in the transition from the primary to the secondary core, repeating the relevant data in (65a,b) and adding an instance where a numeral follows the adjective nye (65c):

(65)	a.	Ny-e	Grei-Ø	Kafeteria		(Norwegian)
		New-wk	Nice-st	Cafeteria
		‘New Nice Cafeteria’
	b.	Ny-e	Elses	Blomster
		New-wk	Else’s	Flowers
		‘New Else’s Flowers’
	c.	Ny-e	Fire	Vegger	[store]
		New-wk	Four	Walls
		‘New Four Walls’

In addition, recall the different interpretative effects with nye from above, which we also used to motivate the two different cores. To repeat then, it is clear that the primary core in Norwegian involves a DP.

We can propose now that if nye is part of the primary core of the Lexical Entry, it is an ordinary adjective in Spec,AgrP followed by a regular lower structure (Subsect. 4.3). Being part of a regular structure, it receives a—what we have called—primary interpretation. If nye is part of the secondary core of the Lexical Entry, it will get a secondary interpretation. Specifically, unlike in the first case, we suggest that the second type of nye is an operator-type element (OP). Informally, this element takes an existing PN and returns a new one.²⁸ Structurally, it is a head in a functional phrase (FP). We propose that the head nye selects a proprial DP. Following the suggestion by a reviewer, we formulate the following Lexical Entry involving a selectional restriction: nye_OP: ____ DP_[+proper].²⁹ To be clear, nye has two different lexical entries that involve different interpretations: a regular adjective housed in Spec,ArgP and an operator occurring in a head position.

Structures similar to nye with a primary interpretation were illustrated in Subsect. 4.3. Focusing here on nye with a secondary interpreation, we can illustrate this as in (66). Note that there are two definiteness features, one in the primary core and one in the secondary core of the Vocabulary Entry in (66a). The latter also surfaces in a DP-layer but here outside the primary core as illustrated in (66b):

(66)	a. Ny-e	Elses	Blomster:
	New-wk	Else’s	Flowers
	{↑nye[+p] { blomst[+p], else[+p]; [+pl][+p], [+def][+p]} [+def][+p]}

To be clear, similar to the cases involving the primary interpretations, the presence of the definiteness feature in the secondary core also yields a DP. In other words, all PN, whether they involve primary or secondary cores, project DPs in Norwegian.

Note again that these structures can be referential. Recalling that the definiteness feature and its [+proper] marking are both syntactic in nature, we point out that there is no element in the DP-level that brings about referentiality. If the current discussion is on the right track, then this adds to the discussion of German confirming that referentiality does not originate in the DP-level. As in German, referentiality seems to come from a lower element. Again, we assume that it is due to the referential pointer, here on the operator nye.

If non-restrictive adjectives and non-proprial determiners are present, they are part of the Numeration; that is, the Numeration consists of the primary core, the secondary core—both part of the Lexical Entry, and non-proprial elements. On par with clausal embeddings (e.g., complement clauses; cf. Chomsky’s 2000, 106 notion of subset Lexical Array), we assume that the vocabular elements of the primary core (embedded under nye) are Selected and Merged first. Next, the elements from the secondary core are Merged and finally the non-proprial elements are inserted. We propose that non-restrictive adjectives are Merged differently from restrictive ones. Rather than in Spec,AgrP, we suggest here that they are adjoined to DefP, the left edge of the secondary (referential) core in Norwegian. This is relatable to ideas from Partee (1973, 54), who points out that non-restrictive modifiers combine with the rest of the noun phrase above the definite article (i.e., they combine with referential expressions) but that restrictive ones combine below the definite article (also Jackendoff 1977, 171–82; Roehrs 2009, 97–116; Pfaff 2015, 2017). Non-proprial determiners are inserted in the DP-level. The simplified derivation of (67a) is as in (67b) (the different domains are delineated by arches):³⁰

(67)	a. den	berømt-e	Ny-e	Elses	Blomster
	the	famous-wk	New-wk	Else’s	Flowers
	‘the famous New Else’s Flowers’

Note that due to the movement of the definiteness feature from Def to D, the highest DP is specified for definiteness. This in turn accounts for the weak ending on the lower, non-restrictive adjective. In other words, there is some limited, morphological interaction between the non-proprial adjective and the proprial definiteness feature in (67b).

Finally, we observed that nominal elements in Norwegian are sensitive to definiteness. This is particularly clear with adjective inflections. We followed previous work in assuming that the feature [def] seems to be an integral part of noun phrases in Norwegian. As proposed above, phrasal PN are related to common DPs (e.g., they share the same basic derivation). Thus, if common DPs are sensitive to definiteness, then their phrasal PN counterparts must also have that feature. Now, since these nominals form PN, all elements that are part of the PN must be marked by the feature [+proper]. This yielded a frozen DP-level with phrasal PN in Norwegian. It is well known that nominal elements in the other Scandinavian languages are also sensitive to the presence of definiteness (for general comparative discussion, see, for instance, Julien 2005; Schoorlemmer 2012; Vangsnes 1999). We predict then that these languages also involve a frozen DP-level. As far as we know, this prediction is borne out.

5.2 Structure of German

The above discussion has not revealed any unexpected patterns in German. That German involves regular structures seems to be confirmed by the inflections on adjectives in these cases. An important result of Roehrs and Julien (2014) was that inflections on adjectives in German are sensitive to structure such that preceding determiners only bring about weak inflections in simple, ordinary noun phrases:

(68)	a.	der	groß-e	Indianer		(German)
		the	big-wk	Indian
		‘the big Indian’
	b.	mit	dem	deutsch-en	Chef
		with	the	German-wk	boss
		‘with the German boss’

Complex structures as in (69a,b) fare differently. Here, adjectives must take a strong ending despite the presence of a preceding determiner (examples are adopted from Roehrs 2015, 251, 258)

(69)	a.	der	Indianer	[Groß-e*(r)	Bär]	(German)
		the	Indian	Big-st	Bear
		‘the Indian Big Bear’
	b.	mit	dem	[Deutsch-e(*n)	Bank]-Chef
		with	the	German-st	Bank-boss
		‘with the German Bank-boss’

Note that both (69a,b) involve two nouns. Taking Grimshaw’s (1991) discussion of extended projections into account, this entails two separate nominal structures (the embedded nominal is indicated by brackets above). With the adjective in a different, separate nominal from the preceding definite article, the adjective and determiner are not in the regular structural constellation, and this explains the strong ending on the adjective (for some discussion, see Roehrs 2015).

Now, we have seen that a demonstrative or a possessive determiner can surface in the left periphery of PN (70a). Interestingly, both of these elements trigger a weak ending on the adjective that is part of the PN (70b). In fact, note that all the adjectives following the definite article must have a weak ending, independent of whether they are part of the primary core, the secondary core, or the left periphery (70c):

(70)	a.	diese	/	meine	Deutsche	Bank					(German)
		this	/	my	German	Bank
		‘this / my German Bank’
	b.	Ich	halte	viel	von	dieser	/	meiner	Deutsch-e*(n)	Bank.
		I	think	much	of	this	/	my	German-wk	Bank
		‘I think highly of this / my German Bank.’
	c.	das	berühmt-e(*s)	Neu-e(*s)	Ägyptisch-e(*s)	Museum
		the	famous-wk	New-wk	Egyptian-wk	Museum
		‘the famous New Egyptian Museum’

Now, in contrast to (69a,b), the strings in (70a–c) have only one noun entailing the presence of one (complex) nominal. With the adjectives and definite article in the same nominal, this accounts for the weak ending on all adjectives on par with (68a,b) above.³¹ Overall, this confirms that these phrasal PN have a regular syntactic structure in German, even in the transition from one domain to another.

With that in mind, the different domains can be related to the following structure that involves one nominal. We illustrate (70c) as in (71). There are two differences to Norwegian. First, since a definiteness feature is not present in the Lexical Entries in German, there is no second, lower DP. Second, recall that the status of secondary neu- is not as clear in German as in Norwegian. For concreteness, we will put this element in Spec,FP, which is followed by a regular lower structure. The non-restrictive adjective is Merged outside the secondary domain; here, it is adjoined to FP:

This yields a structure that is similar (but not identical) to Norwegian, the main differences being the status of neu- and the nominal following neu-. Indeed, note that both the primary and secondary domains are not the same: DP and DefP in Norwegian vs. AgrP and FP in German. Recall that domains are defined by the Lexical Entries (and should not be equated with structurally consistent zones across different languages). Related to this, the left periphery is not cross-linguistically fixed—elements to the left of the core(s) are in the periphery.

While the status of neu- as an operator in German is not as clear as in Norwegian, there is some evidence that this element may be in the process of developing into an operator as well. At least in German, this operator-type element seems to have some features of a determiner.³² Comparing (72a) and (72b), recall from above that some speakers marginally allow neu- to be followed by a weak adjective:³³

(72)	a.	Neu-es	Ägyptisch-es	Museum	(German)
		New-st	Egyptian-st	Museum
		‘New Egyptial Museum’
	b.%	Neu-es	Ägyptisch-e	Museum
		New-st	Egyptian-wk	Museum

This is a typical feature of a definite determiner in German. Indeed, such properties have also been identified with so-called adjectival determiners. A well-known example involves folgendes ‘following’. Note that this element can, with some variation, take a strong or weak adjective (Roehrs 2009, 167–168; van de Velde 2011).

(73)	a.%	folgend-es	interessant-es	Beispiel
		following-st	interesting-st	example
		‘the following interesting example’
	b.	folgend-es	interessant-e	Beispiel
		following-st	interesting-wk	example
		‘the following interesting example’

The option of (73b) seems to be developing in (72b).

Finally, if German does not involve a definiteness feature in its phrasal PN including in those extended by neu-, then we predict that German requires a definite article when an extended PN appears in argument position but Norwegian does not. This is borne out:

(74)	a.	*(Die)	Neu-e	Roland-Apotheke	ist	aber	teuer!	[store]	(Ge)
		the	New-wk	Roland-Pharmacy	is	prt	expensive
		‘The New Roland Pharmacy is quite expensive.’
	b.	Ny-e	Amore Pizza	er	ganske	dyr.		[restaurant]	(Nw)
		New-wk	Amore Pizza	is	quite	expensive
		‘The New Amore Pizza is quite expensive.’

We close this section with another prediction and some remarks about the possible relatedness of phrasal PN to idioms.

5.3 Some Further Considerations

We have seen that in the transition from the primary to the secondary core, there are some unexpected patterns in Norwegian. With the DP frozen in the latter language, we predict that in the absence of a secondary core, there should also be some unexpected pattern in the transition from the primary core to the left periphery in Norwegian but not German. This is indeed borne out.

When a possessive in Norwegian is followed by a phrasal PN, that PN can start in a strong adjective (75a). With ordinary DPs, the adjective must show a weak ending (75b):

(75)	a.	Vårt	Ny-tt	Hjem	er	det	beste	i	verden.	[company]	(Nw)
		our	New-st	Home	is	the	best	in	world
		‘Our New Home is the best in the world.’
	b.	vårt	ny-e	hjem
		our	new-wk	home
		‘our new home’

In contrast, possessives in German combine with a phrasal PN in the expected, regular way:

(76)	a.	unser	Deutsch-es	Ärzteblatt	[magazine]	(De)
		our	German-st	Physician.paper
		‘our German Physician Paper’
	b.	unser	deutsch-es	Auto
		our	German-st	car
		‘our German car’

With Norwegian PN involving DPs, the question arises as to where the preceding possessive in (75a) is located.

In more tentative terms, there seem to be two options: (i) either the possessive itself is in the DP-level, and there is a null linking element that takes the PN, itself a DP, as a complement (cf. the structure in (67b)), or (ii) the possessive is in a position higher than the DP-level. There seems to be evidence that the second option is, at least in principle, possible. While the data in Norwegian are not as clear, Danish, a language closely related to Norwegian, shows that there is a position above the DP proper (see also Roehrs 2019).

To begin, note that Danish does not involve a suffixal article if a demonstrative is present (77a). Furthermore, the free-standing article is syntactically optional if a demonstrative and an adjective are present (77b) (the data are from Julien 2005, 113):

(77)	a.	dette	hus			(Danish)
		this	house
		‘this house’
	b.	dette	(det)	høje	hus
		this	the	high	house
		‘this tall house’

Now, with nouns involving human(-like) reference, third-person pronouns can replace the demonstrative above. For such cases, Johannessen (2008, 173–74) points out that the noun must have a suffixal determiner (78a) and that the free-standing determiner is obligatory when an adjective is present (78b) (the data are from Strahan 2008, 212–13):

(78)	a.	ham	mand-en			(Danish)
		him	man-def
		‘that guy’
	b.	ham	den	store	mand
		him	the	big	man
		‘that big guy’

Both Johannessen (2008) and Strahan (2008) analyze the pronouns in (78) as a kind of demonstrative. They conclude that these demonstratives are followed by regular DPs. If so, then we have some evidence that there is a position above the DP-level. We will tentatively assume that possessives such as in (75a) can also be in that position. We turn to the possible relatedness of phrasal PN to idioms.

Bruening (2010) provides a detailed account of idiom formation. For idioms such as (79), he formulates the components in (80) to (82) (p. 532):³⁴

(79)	a.	throw NP a bone
	b.	give NP a wide berth

(80)	The Principle of Idiomatic Interpretation
	X and Y may be interpreted idiomatically only if X selects Y.

(81)	Constraint on Idiomatic Interpretation
	If X selects a lexical category Y, and X and Y are interpreted idiomatically, all of the selected arguments of Y must be interpreted as part of the idiom that includes X and Y.

(82)	Lexical categories are V, N, A, Adv.

Most importantly for current purposes, non-selected items like determiners, certain cases of adjectives, and possessors may vary freely within idioms (83a-c) (examples taken from Bruening 2010, 533):

(83)	a.	pull some discreet / a few / yet more strings
	b.	kick the filthy habit
	c.	lose X’s cool

Furthermore, as already seen in (79), some idioms have open slots. Moreover, some of these idioms can undergo the dative alternation (84a,b) and passivization (84c) including changes in the verbal tense (examples taken from Bruening 2010, 539, 541; see also Jackendoff 1997, Chap. 7):³⁵

(84)	a.	read NP the riot act
	b.	read the riot act to NP
	c.	The riot act was read to the protesters.

As discussed in the current paper, this is very different from phrasal PN—the latter are always frozen. Furthermore, as already pointed out by Anderson (2007, 315), PN are referential but idioms are not. Thus, while certain idiosyncratic aspects of idioms are relatable to phrasal PN (e.g., both may have partially non-compositional semantics), it appears that these two types of elements should not receive the same account (although the non-compositionality of idioms might lead us to expect that the latter have special vocabulary entries as well).

6 Conclusion

In this paper, we discussed phrasal PN in German and Norwegian. We followed Roehrs’ (2020) analysis of these types of nominals in German. In conjunction with the operation Proprialization, this proposal advocates for regular derivations. This explains the hybrid properties of phrasal PN. Norwegian exhibits properties that are different from German or not as easily seen, a frozen DP-level and recursivity, respectively. The presence of a definiteness feature marked by [+proper] in the stored Lexical Entries and the operator nye explained those facts in Norwegian. More generally, given that phrasal PN are clearly recursive (in Norwegian) and that proprial articles can and sometimes must be left out with phrasal PN (in German), it seems unlikely that referentiality originates in the DP-level. Rather, we follow Köhnlein (2015) and Roehrs (2020) in that referentiality is due to the referential pointer present on a lower element in the DP.³⁶ Finally, we focused in this paper on a constrastive discussion of phrasal PN in German and Norwegian. The investigation of phrasal PN in other languages should help us inhance our understanding of these constructions. Of particular interest would be other languages with polydefiniteness such as Greek.