1. INTRODUCTION
I will investigate a demonstrative that deserves more than the sparse treatment it has received so far in the literature. This demonstrative is interesting from a semantic, a pragmatic and a syntactic point of view.
While the deictic properties of demonstratives in the Scandinavian languages have generally been thought to be linked to either spatial distance or discourse (Faarlund, Vannebo & Lie Reference Faarlund, Vannebo and Lie1997), I am going to present another kind of demonstrative here. It has the same phonological form as the third-person singular pronouns, and can only be used with nouns and have human (or human-like) reference. What is interesting about them from a deictic perspective, is that their conditions of use are linked to psychological distance. It is also worthwhile to look in more detail at their conditions of use, and investigate the difference between them and other demonstratives and determiners. Sections 3 and 4 focus on various semantico-pragmatic aspects of the psychological demonstrative.
The psychological demonstratives occur in most of the Scandinavian languages, but interestingly, they have different syntactic characteristics: the languages seem to be divided into two groups, depending on whether the demonstrative is part of the DP or not. The syntactic investigation is carried out in section 5.
I will focus on Norwegian, especially the Oslo dialect, in the sections where the topic is not Scandinavian variation. However, my impression is that in the dialects and languages where this demonstrative can be found, its semantics and pragmatics do not differ from those of the Oslo dialect. The majority of the examples are drawn from two Norwegian speech corpora – the NoTa corpus and the TAUS corpusFootnote 1 – and the Danish BySoc corpus, as well as the web (for Swedish and Icelandic).
2. PRELIMINARIES
2.1 The pronominal psychological demonstrative exemplified
The pronominal demonstrative is exemplified below for Norwegian, Swedish, Danish and Icelandic. It does not seem to exist in Faroese, and while it does exist in the Finland-Swedish dialect of Helsingfors (Magnus Brenner, p.c.), it does not exist in the Finland-Swedish dialects of Österbotten (Jan-Ola Östman, p.c.). Further, it is also debatable whether it exists in north Norwegian. Actually, there seems to be dialectal variation in most of the languages, but I will not pursue this here.
(1)
(2)
(3)
(4)
These examples show that in each case, the demonstrative is a pronoun. But notice that the morphological form of the pronoun differs in the languages. This will be discussed in section 5.
2.2 The deictic dimensions of Scandinavian demonstratives
It is known that the Scandinavian demonstratives are deictic at two levels – spatial and discourse – although not all the dialects have the latter. They are exemplified below:
(5)
The demonstrative denne is used for proximal spatial deixis, while the demonstrative den (here used elliptically) is used for distal spatial deixis.
The discourse-level deixis is rarer, but is used in Faroese and in Hedmark, Norway, exemplified here by the Toten dialect. The discourse-level demonstrative is used to refer to things that have been mentioned in the discourse, while the spatial one is used for things that can be pointed at. The contrast is between the spatially distal demonstrative and the discourse-level demonstrative:
(6)
(7)
To summarise, the spatial dimension is thus divided into two values, proximal (denne in Norwegian and hesin in Faroese) and spatial (den in Norwegian and hasin in Faroese) – both with morhological variants. The discourse-level dimension has only a distal value (den in the Norwegian Toten dialect an tann in Faroese). The pronominal psychological demonstrative will be discussed in the next section.
Although it is difficult to ascertain whether a particular word or construction is new or old when it is so obviously part of the spoken language, Johannessen (Reference Johannessen, Solstad, Grønn and Haug2006, Reference Johannessen, Johannessen and Hagen2008) argues that there are reasons to think that it is relatively new, and still spreading.
3. THE PRAGMATIC AND SEMANTIC DIMENSIONS OF THE PRONOMINAL PSYCHOLOGICAL DEMONSTRATIVE
3.1 A psychologically distal demonstrative
We will see in this section that the pronominal demonstrative is distal along a psychological dimension, i.e. it is a psychologically distal demonstrative (PDD). There are four main conditions on its use. I should make it clear that these main conditions on use are based on my own linguistic insights as a native speaker of Norwegian. They have not been discussed in depth elsewhere. Some of the aspects relating to the conditions 1, 2 and 3 below have been discussed in Johannessen (Reference Johannessen, Solstad, Grønn and Haug2006, Reference Johannessen and Arboe2007, Reference Johannessen, Johannessen and Hagen2008), and some partly by Teleman, Hellberg & Andersson (Reference Teleman, Hellberg and Andersson1999:section 31:274), Delsing (Reference Delsing, Vangsnes, Holmberg and Delsing2003:23) and Lie (Reference Lie, Johannessen and Hagen2008). The specificity requirement (condition 4) has been discussed by Dyvik (Reference Dyvik1979), Lødrup (Reference Lødrup1982), Faarlund et al. (Reference Faarlund, Vannebo and Lie1997), Johannessen (Reference Johannessen and Arboe2007, Reference Johannessen, Johannessen and Hagen2008) and Strahan (Reference Strahan2007).
PDD Condition 1: The speaker does not personally know the person referred to.
(8)
(9)
(10)
In (8) the speaker mentions one person by name only, and another by name and the pronominal demonstrative. The former does not have to be presented in a particular way, but the latter is one that the speaker obviously needs to specify his not-so-close relationship to. In (9) the speaker talks about an actress, and the use of the demonstrative shows that the speaker needs to say that it is not a person close to her. In (10) the speaker also marks a distance to the person referred to, a person who was at one point widely reported in the papers because of an act of cannibalism.
PDD Condition 2: The addressee does not personally know the person referred to.
(11)
(12)
(13)
In (11) the speaker knows that the listener (a recording assistant) does not know her friend, and thus uses the pronominal demonstrative to accommodate to the addressee. In (12) the speaker is uncertain as to whether the addressee knows her driving instructor. In (13) the writer writes at a web-forum where she knows that the others do not know her sister personally. Thus, the writer can help the reader when the latter might not know the person referred to.
Delsing (Reference Delsing, Vangsnes, Holmberg and Delsing2003:23) has a similar understanding of the condition for the Swedish pronominal demonstrative: ‘talaren är osäker på om lyssnaren vet vem som avses’ [the speaker is uncertain as to whether the listener knows who is intended]. Teleman et al. (Reference Teleman, Hellberg and Andersson1999:§31:274) also point out that the listener's knowledge of the person referred to can be a condition on the use of this pronominal demonstrative. They say that the speaker and the listener must know of the person referred to, but that this person must not be a close acquaintance or be activated in the discourse. I agree with Teleman et al. that both the speaker and the listener should know of the third person, so that it is at all possible for the listener to identify him or her. This knowledge can be very indirect, though, as in example (14) below, where it is world knowledge that constitutes the way the listener should be able to identify the person referred to.Footnote 2
However, I do not agree with Teleman et al. that the person should not be activated in the discourse. Like Teleman et al., Lie (Reference Lie, Johannessen and Hagen2008) focusses on discourse activation, and says that the main task of the psychological demonstrative is that of bringing a new person into the discourse, and Strahan (Reference Strahan2007) also stresses the discourse activation aspect. Given the perfect acceptability of an utterance like (14b) below, used after a person is introduced and well-established in the discourse, it is clear that discourse activation cannot be a necessary condition for the use of the PDD. A second argument against discourse activation as a factor here comes from examining key contexts for discourse activation. A central way of activating referents in the discourse is by using left dislocation (discussed in Svennevig Reference Svennevig, Svennevig, Sandvik and Vagle1995). If the key function of the PDD were to activate referents in the discourse, then left dislocation would be accompanied by the PDD. But this is not the case. They can co-occur, but the PDD then adds an element of psychological distance, and not discourse activation.Footnote 3
Example (14) shows, on the one hand, how personally distant the listener can be from the person referred to, and, on the other hand, that repeated use of the PDD is possible with a person activated in the discourse.
(14)
Wife: Can you fetch my book at the library?
Husband: OK. How do I do it? I'm never at the library.
(a)
(b)
The wife in this dialogue knows that the husband has no knowledge of the person she is talking about. However, she does know that her husband has general knowledge about libraries (people work there, they are often women, they are there to help the customers), and this is enough to introduce a finite set of librarians, and hence identify one of them, however indirectly. Thus, his slight knowledge of the person referred to makes her use the PDD.
Notice that the wife does not use the PDD when she names the library woman in (14a). The PDD would have been totally unacceptable there. Had the husband known already that the woman the wife was describing was called by that name, the PDD would have been ok, as indeed it is when he utters it. Thus, the PDD needs to be anchored either to a common memory (as in (14b)), or to something already in the utterance, as in (14a), where the preposition phrase anchors it.
PDD Condition 3: The speaker has a negative attitude to the person referred to.
(15)
(16)
In (15) the speaker expresses a negative attitude to the woman she talks about, and in (16) it is the writer's mother who is written about in a derogatory way. It is probably not far from a word being used for general psychological distance to it being used for negative attitutudes.
PDD Condition 4: The person referred to must be specific.
(17)
The PDD must refer to somebody specific, never hypothetical or non-exisiting. This feature will be more thoroughly discussed in section 4.2, where it will be compared with the spatial distal demonstrative.
3.2 The PDD in a wider pragmatic perspective
3.2.1 Psychological deixis
We saw above that the deixis that is known in the Scandinavian languages is spatial deixis and discourse deixis. Cross-linguistically, it is also common to count social deixis, often tied to honorifics (‘such qualities as kinship relations, totemic relations and clan membership’ (Levinson Reference Levinson1983:90). In addition, emotional deixis has been mentioned in the literature. One of the first to notice this kind was Robin Lakoff: ‘Both this and that establish solidarity and imply shared emotions between the speaker and the listener’ (Lakoff Reference Lakoff, LaGaly, Fox and Bruck1974: 345f.).
Our psychological demonstratives are possibly emotional in the sense described by Lakoff, because they can create shared emotions between two speakers in the situation where both refer to a person by the PDD. However, they just as often create distance between the two speakers, as in (18). If speaker 1 knows the third person well, she will refer to this person by that person's first name. However, if speaker 2 does not know this person, speaker 2 will avoid using the first name on its own, because it would indicate a personal knowledge that is not there. Speaker 2 could as an alternative use the first and the second name, but that would be unacceptable. So speaker 2, whenever he has to refer to speaker 1's friend (and cannot use a simple pronoun), will use the only possible tool: the PDD plus the first name or the PDD plus a descriptive DP. Let us illustrate:
(18)
Thus, speaker 2 is in a situation where he has to show his non-personal knowledge of Brad in every utterance, unlike speaker 1. This hardly gives the two speakers a sense of solidarity.
Brown & Levinson (Reference Brown and Levinson1978) write about politeness theory and face-threatening acts, and how we use language to avoid them. Lie (Reference Lie, Johannessen and Hagen2008) suggests that this theory can be used to explain the use of the psychological demonstrative. I agree that this is a possible reason for the use of the PDD, condition 2, where the speaker knows that the addressee does not know the person referred to, and chooses to convey this by using the PDD. In a sense, it can also be used to explain the use in condition 1, in which it is the speaker herself who is not acquainted with the person referred to. First, it gives the adressee immediate knowledge about the speaker's level of personal acquaintance with the third person, which is itself a helpful piece of knowledge for the adressee. Second, in contexts where the adressee knows the third person better than the speaker does, e.g. in a dialogue like that in (18), the PDD makes it possible for the speaker to talk about the third person by that person's first name, without coming across as somebody who is pushing their intimacy onto the adressee and his friend.
Psychological distance has also been used previously in the literature, e.g. by Aksu-Koç & Slobin (Reference Aksu-Koç, Slobin, Chafe and Nichols1986) to account for aspects related to Turkish evidentials and in Suzuki (Reference Satoko2006), to analyse expressions of surprise relating to the Japanese uchi ‘inside’ and soto ‘outside’. Now we know that it also exists in the grammar of the Scandinavian languages.
3.2.2 Whose perspective?
Deixis is a way of anchoring events in the world. Levinson (Reference Levinson1983:63f.) says that deixis is generally (but not invariably) organised in an egocentric way. There is a deictic centre, and the central person is the speaker. Levinson says that there are some exceptions to this, and one of them is that in some languages, demonstratives can be organised around participants other than the speaker. The Philippine language Sarnal is given as an example where demonstratives are divided into four categories, of which two are the same as our PDD: 1) close to the speaker, and 2) close to the addressee.
Examples of non-speaker oriented deixis are usually taken from Latin, Turkish or languages far away from northern Europe both geographically and genetically (see e.g. Levinson Reference Levinson1983). It is therefore nice to see this feature in the grammar of the Scandinavian languages.
3.3 Semantic, pronominal features in the demonstratives
It is debatable whether the pronominal demonstratives are something other than the pronouns themselves. Whether they are regarded as part of the category of demonstratives or simply pronouns used exophorically (without their own reference) is a question of taste. But it is clear that the pronominal demonstratives have inherited their main features from the pronouns. First, their morpho-syntactic features remain:
(19)
(20)
Second, the in-built specificity requirement of pronouns also holds for our pronominal demonstratives:Footnote 4
(21)
(22)
4. A COMPARISON WITH OTHER DETERMINERS
There are several determiners that are competing with the PDD because they are similar either in form or in distribution. Below we shall compare the PDD with the preproprial article as well as with the spatial demonstrative.
4.1 The PDD is different from the preproprial article
Many Scandinavian dialects have a preproprial article that is used with names and close family relations. In many cases, the phonological form of this article is the same as that of the PDD. One could wonder if they should be regarded as the same category. However, there are some important differences (cf. Fretheim & Amfo Reference Fretheim and Amfo2005; Julien Reference Julien2005; Johannessen Reference Johannessen, Solstad, Grønn and Haug2006, Reference Johannessen and Arboe2007, Reference Johannessen, Johannessen and Hagen2008).
i. The preproprial article can be inflected for case in Norwegian (e.g. in the Toten dialect, Faarlund Reference Faarlund2000), while the Norwegian PDD is not – it always has the nominative form. This is the general pattern in all the languages that have both, with the exception of Icelandic.
ii. Although the PDD has the same form as the preproprial article (in some inflectional forms at least), there are important differences. First of all, the PDD has some stress, while the preproprial article is always unstressed, to the extent that it is always cliticized to its neighbouring word.
iii. In many dialects, the form of the preproprial article is quite different from that of the PDD. This is the case in the Oslo dialect, where the forms of the preproprial article are/were a (fem.) and n (masc.), while the PDD are hun or hu (fem.) and han (masc.), as we have seen in many examples of above.
(23)
(24)
iv. While the PDD adds meaning to its noun phrase, i.e. that of psychological distance, this is generally not the case with the preproprial article (Delsing Reference Delsing, Vangsnes, Holmberg and Delsing2003). In north Norwegian dialects, the general trait is that the preproprial article is used with absolutely all names, plus certain family relations like ‘mother’ and ‘father’. In these dialects, the preproprial article offers very little meaning, other than that of signalling that a proper name follows (Iversen Reference Iversen1918). In other dialects, e.g. the Toten dialect, the preproprial article is used only with first names (Faarlund Reference Faarlund2000). Since people are referred to by their first names only when they are personally known to the speaker (and addressee), it could be claimed that there, the preproprial article has a psychologically proximal meaning, i.e. one that signals closeness rather than distance.
v. The preproprial article occurs only with names (incluing family relations), while the PDD occurs with all kinds of nouns denoting humans.
vi. The preproprial article is obligatory with the category it occurs with, while the PDD is optional.
4.2 The PDD has different conditions of use from the spatial demonstrative
Here, we shall look at some differences between the PDD and the spatial demonstrative (SD). We have already seen the main conditions of use for the PDD in section 3, but it is only when we compare it with the spatial demonstrative that we really can see how they differ.
SD Condition 1: Identifying somebody in a geographical distance
Let us start with a couple of sentences with the spatial demonstrative.
(25)
(26)
These two examples are from the NoTa corpus, in which all the recordings involve just two people at a table. In both cases, the person referred to is the project assistant who is present during the recordings. These examples are interesting because they show that it is acceptable to use the spatial demonstrative in front of the people who are being referred to. The PDD would have been inappropriate: if the speakers had used the PDD it would have indicated that they wanted to underline the fact that they did not have personal knowledge of the person referred to. This would have been a rather impolite thing to insist on in front of that person.
The purely identificatory function of the spatial demonstrative is also seen clearly in the sentence pair below, where we test the two demonstratives in front of proper names:
(27)
(28)
In (27) it is quite clear that there are several girls named Fia, and that the speaker refers to one of those many girls.Footnote 5 In (28) there is no indication at all that there exists more than one girl called Fia. Its condition of use is, rather, psychological distance as described in section 3.
SD Condition 2: Identifying non-specific types
Further examples show that the spatial demonstrative can identify non-specific types:
(29)
(30)
Both these sentences refer to types rather than specific tokens (persons). In (29) the speaker describes a general type of situation that annoys her, and in (30) the speaker describes a teaching system where different groups go with different teachers.Footnote 6
SD Condition 3: Predicative
Whether the ambiguity between the demonstrative den and the determiner den is something more than just difference in stress is debatable (see also fn. 6). The examples we have seen with the PDD so far have been clear demonstrative usage, with the PDD modifying a noun, and being optional from a syntactic point of view. However, occasionally, the gendered pronouns can occur as adjectival determiners. Determiners are obligatory with definite adjectivees, usually it is the SD type den which is used, and they are often unstressed. We now face a possible opposition for the adjectival determiners just as we did for the demonstratives. Below are two examples with the SD type of determiner, used predicatively.
(31)
(32)
In (31) the speaker talks about the actor Claes Gill and a role he once had as a priest. In (32) the topic of conversation is a boy called Frode, characterised as being the new favourite.
It is interesting that both these examples would have got a totally different semantics had the pronoun han ‘he’ been used as a determiner instead of den. We exemplify with the latter example. The definite description den nye gullgutten ‘the new favourite’ refers indirectly to an individual, attributively, and is more like a quantifier; it does not have a strong presupposition of familiarity, and is easy to accommodate. The phrase han nye gullgutten refers directly to an individual, referentially, as if it were a pronoun. It has a strong presupposition of familiarity, and is difficult to accommodate. Thus, den and han give very different possibilities with respect to semantic interpretation.
5. THE SYNTAX OF THE PRONOMINAL PSYCHOLOGICAL DEMONSTRATIVE
5.1 Introduction
Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005) both describe a syntactic structure in which the demonstrative is situated in a phrase (SemP and DemP, respectively) above, or externally to, the DP.Footnote 7 While Josefsson primarily writes about Swedish, Julien takes all the Scandinavian languages to be analysable with reference to the same structure. They suggest that the structure of the DP with the demonstrative is as in (33).
(33) [DemPhan [DP den gamle vaktmästeren]] ‘that old caretaker’
In this paper, I will argue that the pronominal psychological demonstratives have different syntactic properties in the Scandinavian languages. It turns out that Icelandic and Norwegian constitute one group, and Swedish and Danish another. For the two former languages the demonstrative is internal to the noun phrase, while for the other two, it is external to it. There is one major difference between the languages that leads to this conclusion, but there are also other facts, as we shall see. The main parameter is whether definite determiners can co-occur with the PDD:
(34) PDD ± definite determiner + adjective + noun
We will look at the four languages with attested PDD in turn and ask about the possible co-occurence of the PDD and preposed determiners, and other questions that will shed light on the syntactic structure.
5.2 Danish
5.2.1 Can the PDD and the preposed definite determiner co-occur?
We start with the main question above: can the PDD and the preposed definite determiner co-occur? The relevant examples are given below.
(35)
(36)
(37)
Danish, like the other mainland Scandinavian languages, needs a preposed determiner whenever a singular noun is modified by an adjective, as in (36). We see in (35) that this preposed determiner remains when the PDD is used. Leaving out the preposed determiner yields an ungrammatical result, (37). We can conclude that in Danish, the PDD and the definite determiner co-occur, which lends support for the DP-external hypothesis of Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005).
5.2.2 Does the PDD affect the rest of the DP?
In Danish, there is a general prohibition against definiteness being expressed more than once. This is something we can use to see to what extent the PDD affects the rest of the DP.
(38) pigen
girl.def
(39) den pige / *pigen
(40) hende *pige / pigen
In (38) we see how an unmodified definite noun has a definiteness suffix. In (39) we see what happens when a preposed determiner is added; the definiteness is now expressed at the preposed determiner, and it would be ungrammatical to retain the definiteness marking on the noun. Interestingly, and surprisingly, when the PDD is used, it does not allow the noun to be indefinite, (40). Thus, it seems that the PDD is outside the DP, where it cannot affect it.
5.2.3 What is the case situation for pronouns and for the PDD?
Pronouns heading their phrases are case-marked in colloquial Danish. Ordinary subjects have nominative case, while accusative case is used with a wider variety of functions: preposision complement, direct object, indirect object, coordination,Footnote 8 predicative, associate of relative clause. Nominative subjects are presented in (41)–(42), while the range of accusative pronouns are exemplified in (43)–(48).
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
While pronouns have case distinctions, the PDD does not.Footnote 9 Its form is invariably the accusative hende ‘her’ and ham ‘him’, including where ordinary pronouns would be nominative, such as with subject (and dislocated subject):
(49)
(50)
(51)
(52)
What we have seen here, then, is that Danish pronouns get case when they are heading their phrases, so that subject pronouns have the nominative form and object pronouns have the accusative form. The PDD does not vary with respect to case. This might also be support for the DP-external view; that case is distributed in the DP proper, but not to its external projections.Footnote 10
5.3 Swedish
5.3.1 Can the PDD and the preposed definite determiner co-occur?
We recall that the PDD can co-occur with the definite determiner in Danish. This is also the case in Swedish:
(53)
The fact that it is possible to have both determiner and the PDD present at the same time, as in (53), lends support for the DP-external hypotehsis suggested for Swedish by Josefsson (Reference Josefsson, van Riemsdijk and Hellan1994, Reference Josefsson and van Riemsdijk1999, Reference Josefsson2005) and Julien (Reference Julien2005).Footnote 11
5.3.2 What is the case situation for pronouns and for the PDD?
Pronouns are normally distinguished with respect to case in Swedish. Subjects have nominative case, while accusative case is used for direct and indirect objects, and preposition complements. Exemplified below are nominative and accusative forms of the two animate pronouns.
(54)
(55)
(56)
(57)
However, it should be mentioned that in the spoken, colloquial language the case contrast between nominative and accusative is neutralised, see examples (58)–(59).Footnote 12
(58)
(59)
However, unlike in Norwegian (see section 5.5 below), accusative forms cannot be used as subjects. In addition, Swedish has pronominal clitics, but they are used as objects.
(60)
(61)
Thus, the case system for Swedish pronoun seems to be rather stable, with nominative forms as subjects and accusative forms as objects. However, the form of the PDD is invariable irrespective of the syntactic function of the DP in which it belongs.Footnote 14 In (62) it is in a subject, and in (63), a direct object.
(62)
(63)
Thus, we can conclude for Swedish, as we did for Danish, with respect to case: the fact that there are case distinctions with the pronouns, but not with the PDD, is compatible with the idea that the PDD is DP-external, in accordance with Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005).
5.3.3 Specificity and agreement
In languages with ‘double definitness’, like Norwegian and Swedish, there may be some interaction between the preposed determiner and the definiteness suffix on the noun. In Norwegian, the PDD is part of DP, and it thus takes part in this interaction. In Swedish, the PDD is DP-external, and we should not think that it would interact with these features. And indeed it does not. Actually, there is no specificity agreement between the preposed determiner and the suffix. In both cases, the preposed definite determiner can be left out. The examples below are from Julien (Reference Julien2005:38), exemplifying non-specific and specific interpreation, respectively:
(64)
(65)
What we see in these examples is that the use of the definiteness suffix is determined by the specificity conditions of the DP, while the preposed determiner is not restricted in the same way. In other words, there is no specificity agreement in Swedish, neither between the PDD and some part of the DP nor within DP.
5.4 A short summary on Danish and Swedish PDDs
In this section we have investigated Danish and Swedish with respect to the overarching question of whether or not the PDD is part of the DP proper. We have looked at whether the preposed definite deteterminer can co-occur with the PDD; whether the PDD affects the rest of the DP syntactically; whether the case system for pronouns is different from the one of the PDD, and whether there is agreement or some other interaction between the PDD and the DP. Based on the results of these considerations, we can conclude that for Danish and Swedish, the DP-external analysis of Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005) is supported.
5.5 Norwegian
5.5.1 Can the PDD and the preposed definite determiner co-occur?
We have seen that the PDD and the definite determiner co-occur in Danish and Swedish. Norwegian, too, has a preposed definite determiner when there is adjective modification, so the question must be asked for this language as well. Below are the relevant phrases.
(66)
(67)
(68)
In (66) we see how the preposed definite determiner is obligatory when a definite noun is modified by an adjective. In (67) we see how the PDD can actually take the place of the definite determiner, yielding a grammatical result. (68) shows that co-occurence of the PDD and the preposed definite determiner is not acceptable in colloquial Norwegian, although it can occasionally be found in archaic, formal language.Footnote 15
We can conclude that in Norwegian, the PDD and other preposed determiners cannot co-occur, they are in fact in in complementary distribution. This sets Norwegian apart from Swedish and Danish. There is thus nothing that forces us to conclude that the PDD should be DP-external. On the contrary, the fact that the PDD and the preposed definite determiner never co-occur indicates that they occupy the same syntactic position.Footnote 16
5.5.2 What is the case situation for pronouns and for the PDD?
We have seen that Danish and Swedish have a case system for their pronouns, while their PDDs are invariant. We shall see that Norwegian is quite different from the other two mainland Scandinavian languages.
The general rule in Bokmål for the third-person singular (3sg) feminine in the written language is that the subject form is hun, while the object form is henne. For Nynorsk, ho is used for both subject and object with henne an alternative object form.
(69)
(70)
For the 3sg masculine, the Bokmål norm says that the subject form is han, which can also be used as the object form, alongside ham. In Nynorsk, too, han can be used as subject and object, but for the latter there is also the alternative honom.
(71)
(72)
We see that there are very few, if any, case distinctions in both written varieties. In Nynorsk, it is possible to have a system completely without subject and object distinctions, while in Bokmål, the masculine pronoun can have case syncretism.
Spoken data are not quite as easy to come by yet, but the easily accessible NoTa corpus shows that there is a tendency not to use the object form. There are 5278 occurrences of the form han, and only 24 of ham, so clearly the former is used as both subject and object forms by the abslute majority of the speakers. The Oslo dialect has two clitic 3sg pronouns that have no written counterpart, the feminine a and the masculine n. These do not differentiate between nominative and accusative case. They are not infrequent in the NoTa corpus: 1497 occurrences of the masculine pronoun n, and 303 occurrences of the feminine pronoun a. This situation adds to the understanding that case is strictly not a feature of 3sg pronouns.
We must look at the feminine pronouns, too, since in the written Bokmål variety these have obligatory case distinctions. It is easy to find examples showing that the written norm is not followed systematically in the spoken language. Below are four examples of the two feminine forms. The first two show the accusative case form used both as subject and as object. The last two show the nominative form used as subject and object.Footnote 17
(73)
(74)
(75)
(76)
The PDD is invariant with respect to case. Its occurrence in subject and object DPs are shown below.
(77)
(78)
Thus, in Norwegian, Danish and Swedish, the PDD is invariant. However, unlike in Danish and Swedish, the case system in the 3sg domain of Norwegian is unclear. It seems reasonable to claim that spoken Norwegian (of which the PDD is part) does not have a DP-internal case system as far as 3sg is concerned. The fact that even the dictionaries that reflect the official norms allow non-distinct case forms for three out of four 3sg pronouns (Nynorsk and Bokmål, masculine and feminine) supports a view that there are virtually no 3sg case distinctions. Thus, the Norwegian situation is compatible with a system in which the PDD is in the same position as pronouns are at other times; neither of them has case. This is in contrast to Swedish and Danish, where DP has case and the DP-external position does not. Thus, the Norwegian case system is compatible with a DP-internal PDD, and nothing suggests an outer DP-layer from a case point of view.
5.5.3 Specificity and agreement
Julien (Reference Julien2005:2.3) says that the definiteness suffix (a filled n position) adds a meaning of specificity to the noun, giving the contrast in (79)–(80). This claim will be countered below.
(79)
(80)
Further, we have seen that the PDD has a specific meaning. Julien has claimed that the D position is unspecific in Norwegian. Given what we have seen so far, it is clear that the claim is wrong for Norwegian, although it may be true for Swedish (which has DP-external PDD).
However, given that it is possible to have either a PDD or an SD determiner, and the noun can have or not have a definiteness suffix, it is posible to set up a fourfold minimal pair to test the claims above. This is done below:
(81)
(82)
(83)
(84)
In (81) we see that the combination of a specific PDD and no definiteness suffix is unacceptable, while (82) is fine. There is, thus, an agreement requirement with respect to specificity between the PDD and the suffix. In (83) the preposed definite determiner indicates nothing about specificity, and there is no definiteness suffix to add specificity either. The overall meaning is one of genericity. In (84) we encounter something else: the preposed definite determiner says nothing about specificity. But the definiteness suffix on the noun gives a special meaning to this DP. (A predicate like er ‘is’ instead of oppfører seg ‘behaves like’ would give an even clearer understanding of the meaning of this DP.) It is not specific in the same way as (82), and not generic like (83), but it instead gives an idea of a presupposed, specific set of which the DP is part. The meaning we get with the constellation of the non-specific, definite determiner and the definiteness suffix is thus one of specificity as covert partitivity in the sense of Enç (Reference Enç1991) and van Geenhoven (Reference van Geenhoven1998).
The fact that there is agreement between the PDD and the noun suffix suggests that they are in the same constituent, i.e. the same DP.
5.5.4 How can there be gender disagreement DP-internally?
If the PDD in Norwegian is indeed DP-internal, we would expect agreement with respect to phi-features such as gender and number within the DP. However, a number of examples show that this is not the case with respect to gender:
(85)
(86)
In order to be able to account for this apparently problematic state of affairs, I draw on Josefsson (Reference Josefsson2005: exx. (24a, b), (11) and (26)), who shows that there is gender disagreement between our PDD and the DP in Swedish:
(87)
(88)
Josefsson (Reference Josefsson2005) explains the lack of gender agreement by arguing that Swedish distinguishes between grammatical and semantic gender, where hon/han ‘she/he’ have semantic gender, and do not have to agree with the otherwise grammatical gender in the DP. I take Josefsson's argument to cover Norwegian as well. The apparent gender disagreement is thus not a problem for the DP-internal PDD in Norwegian, since there are two different kinds of genders, of which only one (the grammatical one) has to agree.
5.6 IcelandicFootnote 18
5.6.1 Can there be co-occurrence of PDD and determiner?
Icelandic does not have a preposed definite determiner, so the question of co-occurence does not arise:
(89)
To find out whether the Icleandic PDD is DP-external or not, we have to consider other parts of the grammar.
5.6.2 What is the case situation for pronouns and for the PDD?
While the mainland PDDs are all case-invariant, the situation is different in Icelandic. Here, the case system is thriving for pronouns as well as for common nouns and proper nouns. Below are some examples of nominative, accusative and dative pronouns:
(90)
(91)
(92)
(93)
The Icelandic case system is fully intact, as we see. Now let us turn to the PDD.
(94)
(95)
(96)
(97)
Footnote 19 The case forms used with the PDD are the same as those used with the other case-inflected nominals. Thus, nothing in the case system suggests that there is a DP-external projection for the PDD.
5.6.3 Does the PDD affect the rest of the DP?
We saw that in Danish definite determiners affect the morphology of the noun, while the PDD does not. We took this to indicate that the PDD was outside the DP. In Icelandic, there is a similar restriction, however, here the PDD in fact does affect the noun. Consider the following examples.
(98)
(99)
Example (98) shows that definiteness on nouns, such as the subject here, is marked by a definiteness suffix. Example (99) (repeated from section 5.6.2 above) shows how, in contrast, there need be no definiteness expressed on the subject when it is modified by a PDD.Footnote 20 This fact supports the view of the previous section: the PDD is in the same DP as the noun, and what happens at one place in the DP affects another part of that DP; i.e. the PDD must be DP-internal.
5.6.4 What is the gender situation with respect to the PDD and the rest of the DP?
The situation with respect to gender agreement in the Icelandic DP is not unlike that of the other Scandinavian languages when it comes to the PDD.
(100)
(101)
In (100) the PDD is feminine, while the noun is masculine. This suggests that there is semantic gender agreement for some categories, as suggested by Josefsson (Reference Josefsson2005). However, in (101) there is gender agreement between the PDD and the noun. And this agreement pertains even if the person referred to is a woman. With these diverging data, we cannot draw any specific conclusions. (However, even if we were able to find homogenous data, it is unclear to me what that might have contributed to the question of DP-externality.)
5.7 Summary of the syntax of the PDD
We have seen that the syntax of the PDD divides the Scandinavian languages into two groups: Danish and Swedish on the one hand, and Norwegian and Icelandic on the other. The first two languages fit into the structure in (102), in accordance with Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005), where the PDD is DP-external. I have argued, however, that there is no reason for the latter two languages to be analysed as DP-external; on the contrary, they are best analysed as DP-internal, as in (103).
(102)
(103)
The tree structure for the two DP-external languages, Danish and Swedish, is given in (104).
(104)
The psychological demonstrative has its own position above DP. I assume the DP to be a phase (with Svenonius Reference Svenonius, Adger, de Cat and Tsoulas2004), which means that there will be no grammatical interaction between the DemP and the DP. For the tree with a DemP position, this means that:
(i) the D position is free to host a preposed definite determiner, thus allowing co-occurrence between demonstrative and determiner (sections 5.2.1 and 5.3.1);
(ii) the psychological demonstrative cannot affect the DP syntactically, thus allowing a definiteness suffix on the Danish noun even when a PDD is present (section 5.2.2);
(iii) there is no possibility of case agreement between the demonstrative and the DP, and it is thus possible for the DP to have full case-marking, while the PDD is invariant (nominative form in Swedish, accusative form in Danish; sections 5.2.3 and 5.3.2);
(iv) there can be no gender agreement between the PDD and the DP, allowing for the PDD to have semantic gender while the DP has grammatical gender (sections 5.2.4 and 5.3.3); and
(v) there can be no specificity agreement between the PDD and the DP, making it possible to have specificity marking on the noun, but non on the PDD (section 5.5.4).
A structure without a DemP position restricts the possibilities. It is illustrated in (105).
(105)
This tree, for a DP-internal PDD, depicts the Norwegian and Icelandic situation, and it means the following:
(i) the psychological demonstrative and the preposed definite determiner can never co-occur, so they will always have complimentary distribution (sections 5.5.1 and 5.6.1);
(ii) if there are regular case distinctions, they will be present via agreement on all parts of the DP, including the PDD, so that the Icelandic PDD, but not the Norwegian PDD, will have case distinctions (sections 5.5.2 and 5.6.2);
(iii) if there is specificity agreement, it will be present via agreement on all relevant parts, including the PDD, so in Norwegian the PDD will never co-occur with a noun without the definiteness suffix (section 5.5.3);
(iv) if there is grammatical gender agreement, it will be present throughout the DP, including the PDD, as it is in Icelandic, but not Norwegian, where the PDD has semantic gender, as in Swedish (sections 5.5.4 and 5.6.4); and
(v) the grammatical properties of the PDD will affect the rest of the DP, and therefore, in Icelandic, the PDD will cause the noun to be bare, without a definiteness suffix (section 5.6.3).
6 Conclusions
In this paper I have investigated the pronominal psychological demonstratives in the Scandinavian languages. We have looked at them from a semantic, a pragmatic and a syntactic point of view. Semantically, they have some of the characteristics of pronouns, such as specificity and of course their number and gender features. The specificity distinguishes them from definite determiners, and also interacts in interesting ways with the definiteness suffix in some languages.
Pragmatically, there are four conditions of use: 1) the speaker does not know the person referred to, 2) the addressee does not know the person referred to, 3) the speaker may have negative attitudes to the person referred to, and 4) the person referred to must be specific. The psychologically distal demonstrative (PDD) is also compared to spatial demonstratives and the preproprial article, and a number of important differences are identified – phonological, semantic, pragmatic and syntactic.
Finally, I have looked at the syntactic characteristics of the psychological demonstratives. Taking the analyses in Josefsson (Reference Josefsson2005) and Julien (Reference Julien2005) as a starting point, I have loked at a range of syntactic data from the four Scandinavian languages languages that have the PDD. The data include, inter alia, possible co-occurrence of PDD and preposed definite determiner, case agreement, gender agreement, specificity agreement, interaction between PDD and other parts of DP. I have come to the conclusion that Josefsson's and Julien's analyses, both of which include a DemP outside the DP, can only apply to two of the languages, namely Danish and Swedish. It turns out that there are no arguments in favour of analysing Norwegian and Icelandic that way – for these two languages, the DP is itself sufficient.
ACKNOWLEDGEMENTS
I would like to thank several people who have discussed this topic with me: Lars-Olof Delsing, Jan Terje Faarlund, Thorstein Fretheim, Elly van Gelderen, Kristin Hagen, Zakaris Hansen, Svein Lie, Arne Martinus Lindstad, Helge Lødrup, Wiltrud Mihatsch, Kjartan Ottósson, Hannes Rieser, Henrik Rosenkvist, Sigga Sigurjónsdóttir, Tania Strahan and Jan Svennevig. A special thank you to Kjell Johan Sæbø who has helped me sort out the semantic terminology. I am also very grateful for the astute comments of the anonymous reviewer and the thorough remarks of the copy-editor, both of which have helped to improve this paper. I would also like to thank the audiences at the Seminar for Theoretical Linguistics, UiO, autumn 2006; the Grammar in Focus workshop in Lund 2006; the Seminar on the Oslo dialect, 2006; the Nordic Microcomparative Syntax (NORMS) Workshop, Tromsø, 2006; The 8th Nordic Dialectologist Conference, Århus, 2006; the panel of Corpus-Based Pragmatics at the 10th International Pragmatics Association Conference, Gothenburg, 2007; and the NORMS Workshop on Binding and Pronouns, Reykjavik, 2007.
