an l 2 perspective on possessives : contrasts and their possible consequences

The present paper presents the contrastive background and the basic objectives of a cross-linguistic research project (POSS) that takes an L2-oriented perspective on possessives in English, Norwegian, German, French and selected Slavic languages. Our paper focuses on L1/L2 pairs involving Norwegian as L2 or L1. Section [1] outlines the rationale behind our project. Themorphosyntactic (‘core’) systems of English, French, German, Norwegian and Russian third possessives are described and compared in section [2] while section [3] draws attention to dimensions of contrasts that fall outside the scope of our project. Section [4] specifically addresses the L2 issue, presenting for selected L1/L2 pairs our basic assumptions concerning challenges to the acquisition of the L2 possessive core system. Section [5] contains a concluding summary. [1] introduct ion : an l2 perspect ive on possess ives Linguistic expressions of possession (in a wide sense) are a fairly well established topic of cross-linguistic research (see among others (Alexiadou 2007; Baron et al. 2001; Börjars et al. 2013; Chappell & McGregor 1996; Coene & D’hulst 2003; Heine 1997; Koptjevskaja-Tamm2002, 2003;Manzelli 1990;McGregor 2009; Zifonun 2005)). To our knowledge, however, in-depth comparisons of related but somewhat different systems of possessives are scarce (but see Drewnowska-Vargáné & Zifonun (2011); Gunkel et al. (2017, B1.5.4); Ramm & Fabricius-Hansen (2012); Zifonun (2005)). Accordingly, little is known as to whether or how morpho-phonological or syntactic similarities and differences between L1 and L2 possessives1 (‘possessive contrasts’) are reflected in native versus non-native acquisition, use and processing of such items, i.e. what role any influence from L1 — so-called transfer (Benati & Angelovska (2016, 31–58); Ellis (2008, 349–402); Jarvis & Pavlenko (2008); [1] We use L2 as a cover term for L2 and foreign languages acquired after L2. In accordance with Zifonun (2005) we subsume under the category ‘possessive’ both possessive determiners like French mon/ma/mes ‘my’ and genitive forms of so-called personal pronouns like English his. [4] cathrine fabricius-hansen, hans petter helland, anneliese pitz Meisel (2000); Odlin (2003)) — could play in this area. In fact, compared to their non-possessive counterparts, possessive pronouns seem to have been strangely neglected in psycholinguistic (L1 and L2) research (but see Marinis (2016)); for relevant L2 studies concerning non-possessive third person anaphoric pronouns (he, she etc.), including reflexives (himself, herself etc.), see e.g. Clahsen & Felser (2006); Felser & Cunnings (2012); Patterson et al. (2014); Roberts et al. (2008); Umesh et al. (2016) and further references therein. An important issue discussed in these and many other publications on L2 acquisition is the division of labor between what may be seen as specific L1 influence (transfer) and general L2 processing effects (so-called general learner effects). Our paper presents the cross-linguistic background and the basic objectives of L2-oriented research on adnominal possessives that may shed new light on this issue. Some preliminary empirical results are presented by Pitz et al. (2017), Helland (2017) and Behrens (2017).2 From a cognitive point of view, possessives seem more complicated than ordinary pronouns due to the fact that they are not only anaphoric (third person alone) or deictic but at the same time relational expressions: An anaphoric adnominal third person possessive like an ordinary anaphoric pronoun demands an antecedent DP; as a determiner or modifier (see section [3.1]) within a DP, it anchors the referent of its host DP to the referent of the antecedent DP by a relation of possession in a more or less broad sense (see references above), where the antecedent DP denotes the ‘owner’ (the possessor) and the host DP the ‘owned’ entity (the possessum). Thus, processing an anaphoric possessive in a given context involves the following subtasks: (i) identifying (the lexical head of) its host DP, (ii) finding a proper antecedent (i.e. anaphoric resolution), and (iii) using that and the relationalmeaning of the possessive to establish a referent for the host DP. In the case of first and second person and deictically used third person possessives, the possessor is provided by the non-linguistic context. In this paper we are concerned with English, German, Norwegian (bokmål), French and Russian, which all distinguish formally between first, second and third person possessives, exhibiting two or more formally different words (lexical items) of the last category (Faarlund et al. (1997, 203–208); Huddleston & Pullum (2002, [2] In addition to the authors of this paper the following persons participate in the (POSS) project, which is part of a broader research project on Language as Product and Process under the leadership of Bergljot Behrens: Bergljot Behrens (U Oslo), Oliver Bott (U Tübingen), Torgrim Solstad (ZAS, Berlin), Barbara Mertins (TU Dortmund), and Katarzyna Stachowiak (U Warsaw). We thank Begljot Behrens, Oliver Bott, Katarzyna Stachowiak, Hildegunn Dirdal (U Oslo) and an anonymous reviewer for very useful comments on earlier versions of this paper. OSLa volume 9(2), 2017 an l2 perspective on possessives: contrasts and their possible consequences [5] Language First person Second person Third person (possessor = speaker) (possessor = addressee) (poss. ≠ sp. and addr.) English my; our your his, her, its; their German mein*; unser* dein*; euer*; Ihr* sein*; ihr* Norwegian min/mi/mitt/mine; vår/vårt/våre din/di/ditt/dine; deres sin/si/sitt/sine; hans, hennes, dens, dets; deres French mon/ma/mes; notre/nos ton/ta/tes; votre/vos son/sa/ses; leur, leurs Russian moj*; nash* tvoj*; vash* ego, eë; ich } {{ }

The present paper presents the contrastive background and the basic objectives of a cross-linguistic research project (POSS) that takes an L2-oriented perspective on possessives in English, Norwegian, German, French and selected Slavic languages.Our paper focuses on L1/L2 pairs involving Norwegian as L2 or L1.Section [1] outlines the rationale behind our project.The morphosyntactic ('core') systems of English, French, German, Norwegian and Russian third possessives are described and compared in section [2] while section [3] draws attention to dimensions of contrasts that fall outside the scope of our project.Section [4] specifically addresses the L2 issue, presenting for selected L1/L2 pairs our basic assumptions concerning challenges to the acquisition of the L2 possessive core system.Section [5] contains a concluding summary.
[1] i n t r o d u c t i o n : a n l 2 p e r s p e c t i v e o n p o s s e s s i v e s Linguistic expressions of possession (in a wide sense) are a fairly well established topic of cross-linguistic research (see among others (Alexiadou 2007;Baron et al. 2001;Börjars et al. 2013;Chappell & McGregor 1996;Coene & D'hulst 2003;Heine 1997;Koptjevskaja-Tamm 2002, 2003;Manzelli 1990;McGregor 2009;Zifonun 2005)).
[4] cathrine fabricius-hansen, hans petter helland, anneliese pitz Meisel (2000); Odlin (2003)) -could play in this area.In fact, compared to their non-possessive counterparts, possessive pronouns seem to have been strangely neglected in psycholinguistic (L1 and L2) research (but see Marinis (2016)); for relevant L2 studies concerning non-possessive third person anaphoric pronouns (he, she etc.), including reflexives (himself, herself etc.), see e.g.Clahsen & Felser (2006); Felser & Cunnings (2012); Patterson et al. (2014); Roberts et al. (2008); Umesh et al. (2016) and further references therein.An important issue discussed in these and many other publications on L2 acquisition is the division of labor between what may be seen as specific L1 influence (transfer) and general L2 processing effects (so-called general learner effects).Our paper presents the cross-linguistic background and the basic objectives of L2-oriented research on adnominal possessives that may shed new light on this issue.Some preliminary empirical results are presented by Pitz et al. (2017), Helland (2017) and Behrens (2017). 2  From a cognitive point of view, possessives seem more complicated than ordinary pronouns due to the fact that they are not only anaphoric (third person alone) or deictic but at the same time relational expressions: An anaphoric adnominal third person possessive like an ordinary anaphoric pronoun demands an antecedent DP; as a determiner or modifier (see section [3.1]) within a DP, it anchors the referent of its host DP to the referent of the antecedent DP by a relation of possession in a more or less broad sense (see references above), where the antecedent DP denotes the 'owner' (the possessor) and the host DP the 'owned' entity (the possessum).Thus, processing an anaphoric possessive in a given context involves the following subtasks: (i) identifying (the lexical head of) its host DP, (ii) finding a proper antecedent (i.e.anaphoric resolution), and (iii) using that and the relational meaning of the possessive to establish a referent for the host DP.
In the case of first and second person and deictically used third person possessives, the possessor is provided by the non-linguistic context.
In this paper we are concerned with English, German, Norwegian (bokmål), French and Russian, which all distinguish formally between first, second and third person possessives, exhibiting two or more formally different words (lexical items) of the last category (Faarlund et al. (1997, 203-208); Huddleston & Pullum (2002, [2] In addition to the authors of this paper the following persons participate in the (POSS) project, which is part of a broader research project on Language as Product and Process under the leadership of Bergljot Behrens: Bergljot Behrens (U Oslo), Oliver Bott (U Tübingen), Torgrim Solstad (ZAS, Berlin), Barbara Mertins (TU Dortmund), and Katarzyna Stachowiak (U Warsaw).We thank Begljot Behrens, Oliver Bott, Katarzyna Stachowiak, Hildegunn Dirdal (U Oslo) and an anonymous reviewer for very useful comments on earlier versions of this paper.
an l2 perspective on possessives: contrasts and their possible consequences 470ff); Riegel et al. (2009, 288-290); Timberlake (2004, 240-256); Zifonun et al. (1997, 40f)).For convenience they are listed in table 1.The starred items in the German and the Russian columns represent sets of inflected forms: mein, meine, meinen, meinem, meines, meiner; dein, deine etc. (see section [2.3] and [2.5]); for the other languages all possible word forms are listed (for details see section [2]).In what follows, we concentrate on third person adnominal possessives. 3 The choice between the different third person options is determined in part by properties of the antecedent DP (or the entity it refers to, i.e. the possessor).From a processing (comprehension) point of view, this means that a possessive provides grammatical or semantic cues governing the search for a suitable antecedent. 4 However, the hierarchy and type of cues vary somewhat across languages, with the result that even phonologically similar and genetically related items like the possessives beginning with s-in German, Norwegian, French and Russian (henceforth: s-possessives; boldface in table 1) are cued differently ('false possessive friends', see section [4.6]).
In addition, the form of the possessive may vary by inflection according to morpho-syntactic features of its host DP, i.e. by so-called grammatical agreement, as is the case with the German and French possessives, with Norwegian sin/si/sitt/sine, and with Russian svoj*; or it may be independent in this respect like [3] Different from the Norwegian third person reflexive possessive si*, the Russian reflexive possessive svoj* is 'impersonal', i.e. neutral with respect to the distinction between first, second and third person possessor (see section [2.5]).
[6] cathrine fabricius-hansen, hans petter helland, anneliese pitz the other possessives in Norwegian and Russian, and the English possessives as well.In the former case, then, processing a possessive involves checking for two sets of features: those relating to the antecedent and those pertaining to the head noun.For prenominal possessives this normally means 'looking' both to the left and to the right.A preliminary comparison of English and French may illustrate the two types of contrasts and their intricate effects (see section [2] for more details).In English, the choice of possessive depends on the antecedent (referent) alone, as illustrated in (1).
( In French, the choice between son/sa/ses on the one hand and leur/leurs on the other hand likewise is determined by the antecedent (singular versus plural), corresponding to his/her versus their, cf.(2).As for son (or sa) versus ses, however, it is the grammatical gender and number of the head noun (chien 'dog': masculine singular, chiens: plural) alone that counts; and likewise for leur versus leurs.That is, while his and her unambiguously demand a male (Peter) and a female (Anna) antecedent referent respectively, son/sa and ses are neutral in this respect: they may have either Jean or Anna as their antecedent, as illustrated in (2) for son and ses.And in contrast to English, French possessives, like adjectival modifiers, are marked for possessum number and -if singular -also gender, agreeing with their head nouns.In other words: son, sa and ses are different inflectional forms of one 'stem' or possessive lexical item, as are leur and leurs; and the two stems are marked for possessor singular and plural, respectively, while their different forms are possessum-dependent, agreeing with the head noun with respect to gender and number.The French possessives are not cued for possessor gender, whether natural (as in English) or grammatical (as in German); se section Viewed from the perspective of L2 acquisition, this complicated interplay between differences and similarities gives rise to the following general assumptions: (i) Achieving native-like fluency in the use and processing of L2 adnominal possessives (in the languages we are concerned with) is a task of varying complexity, depending in part on the degree of isomorphism between the possessive core systems of the specific L1/L2 pair involved. 5 [5] This, of course, is a theoretical simplification.Transfer in L2 production and comprehension can occur from L1, but also from other foreign languages learnt before or along with the actual L2 (Westergaard et al. 2016), and, according to generative grammar, even from Universal Grammar (UG) (cf.Eide (2015)).
OSLa volume 9(2), 2017 an l2 perspective on possessives: contrasts and their possible consequences [7] (ii) For a given L1/L2 pair, some possessive contrasts may be more fundamental than others, impeding L2 performance of even quite advanced L2 learners.
Testing these and related hypotheses, to be refined somewhat in section [4], is the main direct objective of our POSS project.At a more general level, our investigations hopefully will contribute to the ongoing theoretical discussion concerning the role of transfer from L1 in L2 processing.
Our paper is structured as follows: Section [2] gives a contrastive overview of the (morphosyntactic) core systems of third person adnominal possessives in English, German, Norwegian, French and Russian, with a view to other Scandinavian and Slavic languages.In section [3], we briefly comment on additional dimensions of contrast that are also highly relevant from a L2 perspective but which have to be neglected in the present context.Our approach in these two descriptive sections leans partly on a model of comparison developed within the project Grammatik des Deutschen im europäischen Vergleich 'The Grammar of German in European Comparison' (Gunkel et al. 2017)  6 and applied to possessive pronouns by Zifonun (2005) and Gunkel et al. (2017, B1.5.4) (see section [2]).In section [4], we outline for selected L1/L2 pairs what we, in view of the core systems described in section [2], take to be major obstacles to native-like L2 proficiency in the production and comprehension of possessives. 7Section [5] concludes by outlining the way forward for the POSS project.
A final terminological note: In what follows, we shall use the term possessor not only for the entity referred to by the antecedent DP (see above) but also for the linguistic expression (i.e. the antecedent DP) itself, when necessary specifying the intended meaning in a proper way.In a similar vein, the term possessum, introduced above for the entity denoted by the host DP (i.e. the entity that is identified as 'belonging to' the possessor), unless otherwise indicated will refer to the nominal head (dog etc.) of the host DP.Accordingly, grammatical and semantic properties of a possessive that are determined by the antecedent (referent) will be called possessor-/antecedent-related while grammatical (inflectional) features triggered by agreement with the head noun are possessum-related.
[2] c o n t r a s t i n g m o r p h o s y n ta c t i c ( c o r e ) s y s t e m s o f p o s s e s s i v e s [2.1] Preliminaries As mentioned in the previous section, in view of their twofold function adnominal possessives must contain cues to identify the possessor and have the means to anchor the possessum to this entity.It is with respect to marking these relations that the languages exhibit differences.To describe and contrast the systems, [6] See also http://www1.ids-mannheim.de/gra/abgeschlosseneprojekte/gde.html.

[7]
As far as Norwegian versus German and French is concerned, see (Pitz et al. 2017) and Helland (2017) for more detailed discussions. [8] cathrine fabricius-hansen, hans petter helland, anneliese pitz we will make use of parameters presented in Zifonun (2005), Gunkel et al. (2017, B1.5.4) to account for the differing dimensions of the possessive systems across languages. 8We shall concentrate on the parameters that are of relevance for the languages to be considered here, that is, English (En), French (Fr), German (Ge), Norwegian (No) and Russian (Ru).Importantly, as stated above, we will be concerned with third person possessives alone.
The parameters differentiating between the core systems of the languages under consideration are the following: (i) the categorial (part of speech) status of the possessive, i.e. the question whether it may be classified as an inflectional (genitive) form of the personal pronoun or whether it constitutes a lexical item of its own, being inflected like determiners or adjectives; (ii) the morphological properties or categories establishing the relation to the possessor and the possessum, such as person, number and gender; and (iii) reflexivity, i.e. the question whether the language distinguishes formally between reflexive and non-reflexive possessor relations, where reflexive means that the antecedent of the possessive has to be found (as binder) in a local syntactic domain, like the antecedent of non-possessive third person reflexive pronouns (e.g.Norwegian seg, German sich).
To account for reflexivity, one usually turns to Binding Theory (BT) (Chomsky 1981).Reflexive uses obey principle A of BT, which states that 'anaphors', i.e. reflexives, must be locally bound within their binding domain.In technical terms, the binding domain is the smallest clause that contains the possessive DP and its co-indexed antecedent, generally a higher subject.This means that the reflexive possessive in (3) is (co-)referentially dependent on a nominal element that appears as the subject of the clause (m: masculine, r: reflexive). 9,10  (3) Han i he fant found igjen again bilen car.def.m.sg sin i /sin i poss.r.m.sg/poss.r.m.sg

bil. car
Reflexive and non-reflexive possessives are in complementary distribution.For a language like Norwegian this contrast is formally marked: The non-reflexive -more precisely: irreflexive -possessive hans in (4) cannot be bound by the subject pronoun, i.e. ( 4) is deviant under the co-referential reading indicated by [8] The languages she investigates are German, French, English, Italian, Polish and Hungarian (with a view to the Scandinavian languages, Dutch and Spanish).For a complete list of the parameters, see Zifonun (2005) and Gunkel et al. (2017).[9] In the examples, co-referentiality is marked by co-indexation.In its standard formulation (see Chomsky (1981, 183-230)), the binding domain is defined with respect to C-command: A c-commands B iff (i) A does not dominate B and B does not dominate A. (ii) The first branching node dominating A also dominates B.
The possessives in (3) are thus both c-commanded and co-indexed by their antecedents.[10] The Norwegian possessive may occur in post-head position or in the prenominal position of a determiner (Faarlund et al. 1997, 263ff).In the former case the head noun has a definite suffix (-en in the example) in (3), see section [3.3]).
an l2 perspective on possessives: contrasts and their possible consequences [9] the subscripts (ir: irreflexive).(Timberlake 2004, 240-256).The encoding of possessor person, however, will not be discussed in this paper since we are concerned with third person possessives only.In addition to these obligatory categories, properties such as grammatical gender/natural gender or animacy of the third person singular possessor may be expressed if the language makes this distinction in the nominal domain.In the cases where the possessive can be analyzed as a genitive of the personal pronoun, additional inflectional morphology is likely to be precluded (Zifonun 2005, 64).
[2.2] English The English possessive system is the most straightforward one.The possessives his/her/its/their can be analyzed as genitive forms of the third person pronouns he/she/it/they Huddleston & Pullum (2002, 470ff), hence we do not expect possessum-related morphological markers (see Zifonun (2005, 64)).As personal pronouns distinguish between natural gender (male/female) and -human in the singular, the singular possessives will also express these properties of the possessor; cf. ( 7)-( 10).
[11] The expressions his, her, their and its occurring in glosses should be understood as abbreviations for poss.m.sg, poss.f.sg, poss.pl and poss.nonhum.sgrespectively.
(7) Reflexive use Anna i /Henry j loves her i/*j /his j/*i hat/cars.
(8) Non-reflexive use Anna i /Henry j wore a hat/gloves.Her i/*j /His j/*i hat was green./Heri/*j /His j/*i gloves were green.
(9) Non-reflexive use Anna wore [a hat] i .Henry didn't like its i colour. (10) Reflexive use [Anna and Henry] i love their i car/cars.
(11) Non-reflexive use [Anna and Henry] i have a car/two cars.Their i car is red./Theiri cars are red.
The properties of the English possessive system can be represented as in figure 1.As indicated by the highest node, the choice of the possessive item depends exclusively on possessor properties, foremost on the possessor number.In the singular, a further possessor-related feature ±human comes into play, providing just one form for -human (its), while for +human, the possessives vary according to natural gender (male/female) of the possessor. 12 [12] Gray background: grammatical features; orange/yellow background: purely semantic features.
an l2 perspective on possessives: contrasts and their possible consequences In French, the possessive represents a determiner-like part of speech (Helland 2006, 155-158).The third person possessor number is expressed by stem variation (singular son/sa/ses, plural leur) while the gender of the possessor is not marked. 13 As mentioned in section [1], the possessive in French agrees with the possessum in (singular gender and) number.French distinguishes between masculine and feminine gender nouns, and the possessive exhibits two morphological forms, son (masc.)and sa (fem.)expressing this distinction in addition to possessum number (sg.).Hence, with a singular possessum we get son, sa or leur, with a plural possessum ses or leurs, that is one common plural form for both possessum genders.Thus, possessor number is expressed by stem variation, while possessum number is expressed by inflection.Possessum gender is marked only in the possessor singular; cf.examples ( 12)-( 13) and ( 14)-( 15), which correspond in spirit to the English examples ( 7)-( 8) and ( 10)-( 11), respectively.Where relevant, nouns are annotated for gender and number (m: masc.sg., f: fem.sg., pl: plural).( 12 A note on the historical development might be in order.In the third person singular and plural, classical Latin had competing forms for non-reflexive and reflexive uses, eius/suus (third masc.sg.) and eorum/sui (third masc.pl.) (Peteghem 2012).In the evolution from Latin to French, the reflexive variant suus survived in the third person possessor singular, leading to French son/sien.In the third person plural however, the reflexive form sui, which competed with the genitive paradigm (eius/eorum), was replaced by French leur.This meant that French developed (historically) reflexive possessives (son/sien) and (historically) non-reflexive possessives (leur) with both reflexive and non-reflexive uses. [12] cathrine fabricius-hansen, hans petter helland, anneliese pitz ines the stem of the possessive.The remaining part of the system expresses properties of the possessum (agreement).The first distinction made by the possessumrelated morphology is between singular and plural both for singular and for plural possessors.For the singular possessum, the possessive has a masculine and a feminine form.
[2.4] German In German, too, the possessives belong to the category of inflected determiners.The possessor is identified by stem variation of the possessive: sein* for both masculine and neuter singular antecedents, ihr* for feminine singular and for plural antecedents.That is, the possessive ihr* is ambiguous between possessor singular feminine and plural, and sein* between possessor masculine and neuter singular.
Similar to other determiners, the possessive in German is subject to general agreement requirements within the DP.These conditions include gender and number agreement with the head noun, in addition to agreement in morphological an l2 perspective on possessives: contrasts and their possible consequences [13] case marking; cf. ( 16)-( 20), which in principle are comparable to ( 7)-( 11 Figure 3 illustrates the properties of the German third possessive system.The stem is dependent on possessor number and gender (if singular).As to the possessumrelated features, the possessive inflects according to gender, number and morphological case (n: nominative, a: accusative, g: genitive, d: dative) of the host DP.
[2.5] Norwegian (and other Scandinavian languages) As mentioned in section [2.1], Norwegian distinguishes between reflexive and irreflexive possessives.The reflexive possessive si* demands a third person antecedent but exhibits no stem variation relating to (other) properties of the possessor. 14 It agrees in gender and number with the possessum (sg.masc.sin, fem.si, neut.sitt, pl.sine), as illustrated in ( 21) and ( 22).(Since Norwegian bokmål is developing into a two-gender language, feminine being 'suppressed' by masculine gender, we use the term 'common gender' for the latter.This is the term used in relation to standard Danish and Swedish, which are genuine two-gender languages.)Like the English possessives, the irreflexive Norwegian possessives can be analyzed as genitive forms of third person personal pronouns han 'he', hun 'she', den/det 'it', de 'they'.Here, the natural gender of the possessor (male vs. female), grammatical gender (common -or masc./fem.-vs.neuter) and the feature ±human play a role.Hence, we have +human sg.male possessive hans vs. female hennes, -human sg.comm.dens vs. neuter dets and, more straightforward, the plural possessive deres, which is unspecified in other possessor-related respects.As expected, these forms do not inflect.Being irreflexive, they cannot refer to the subject of the clause (see section [3.1] for a description of binding conditions); cf. ( 23 Since alternatives to possessives tend to be preferred with nonhuman antecedents dens and dets 'its' are rather infrequent, in particular in spoken language.
distinction in the hierarchy is between reflexive and irreflexive possessives, where the reflexive has its own paradigm relating it to properties of the possessum while the irreflexive is a genitive form which reflects possessor properties alone. 16 Norwegian shares the fundamental distinction between reflexive and irreflexive possessives with the other Scandinavian languages although there are differences in the details.Thus, the Danish reflexive possessive sin/sit/sine demands a singular antecedent while deres 'their' refers to a plural possessor independently of binding conditions; i.e. different from Norwegian, Danish deres is neutral with respect to reflexivity. 17 [2.6] Russian (and other Slavic languages) The Russian 18 system is similar to the Norwegian system in distinguishing reflexive and irreflexive possessives.Also, the latter can be analyzed as genitive forms of the third person personal pronouns, differentiating possessor gender and number; as in German, however, there is a syncretism between the sg.masculine and the sg.neuter form: masc./neut.ego, fem.eë, pl.ich.
While the irreflexives lack possessum-related features, the reflexive possessives are inflected for gender, number and morphological case in agreement with the possessum.
Consequently, svoju in (28) can only refer to the clause subject whereas ego and eë in (29) must find their antecedent in the preceding context; and likewise for [16] Green background: syntactic feature of possessor.[17] This means that possessor number has a higher position than reflexivity in the Danish hierarchy of relevant dimensions.[18] In the present paper, we use Russian as a representative of Slavic languages exhibiting the distinction between irreflexive and reflexive possessives -whatever the latter are called in the grammatical tradition of each specific language.In practice, our project will be concerned with Czech and Polish rather than Russian.
( This means that the Russian reflexive possessive is not cued for any inherent (semantic or grammatical) properties of the possessor, in contrast to the possessives in the languages considered so far; it only marks the structural position or syntactic function of the possessor.The Russian third person core system -including the 'impersonal' svoj* -is summarized in figure 5. Figure 6 shows the inflection of the reflexive svoj* (n: nominative, a: accusative, g: genitive, d: dative, i: instrumental, p: prepositional case).
[2.7] Summary of contrasts In this section we have described the (third person) possessive systems of English, French, German, Norwegian and Russian in terms of possessor-and possessumrelated properties.First, quite generally, the possessor number determines the stem of the possessive, whether this is a genitive form as in English, Norwegian (irreflexives) and Russian (irreflexives) or has its own determiner-like inflectional paradigm as in French and German.The Norwegian and Russian reflexive possessives, however, are exceptions in this respect, being underspecified with respect to possessor number.
Next, as a general property, the head noun of the host DP -the possessumby agreement determines the inflection of the possessive unless the latter is a genitive form prohibiting further morphological marking.Hence the (sub)systems of genitive forms (English possessives, Norwegian and Russian irreflexive possessives) exhibit possessor features only.They do, however, express more intricate semantic features of the possessor such as ±human (English and Norwegian irreflexives hans/hennes vs. dens/dets), natural gender (English his vs. her and Norwegian hans vs. hennes), features which are left unspecified or underspecified in the other (sub)systems.The following summarizes the main contrasts concerning (lack of) specification in the various dimensions.
(i) The non-Slavic languages except Norwegian are underspecified with respect to reflexivity, i.e. local versus nonlocal binding. 19  (ii) English possessives express semantic properties of the possessor such as ± human and natural gender while possessum-related properties are left unspecified.
(iii) French possessives are underspecified with respect to all possessor properties except for number.
(iv) German exhibits idiosyncratic ambiguities with respect to possessor gender and number, presenting a syncretism between singular masc.and neuter (sein*) on the one hand and between singular fem.and plural (ihr*) on the other hand.
(v) The Norwegian reflexive subsystem is underspecified with respect to possessor properties other than (local) binding but specified for possessum number and grammatical gender (in the singular).In the irreflexive subsystem, semantic possessor properties such as ±human and natural gender are distinguished, but possessum-related features are not marked.
Viewed from a comprehension perspective, then, it appears that the languages with reflexivity permit the best identification of the possessor.Although the [19] English, German and French do have means to avoid corresponding referential ambiguities in natural discourse, e.g. the demonstrative genitive forms dessen (masc./neut.sg.) and deren (sg.fem./plur.) in German, which preferably take a non-subject antecedent (see Bosch & Umbach (2007)).Demonstrative alternatives tend to be stylistically marked, however, and will not be considered here.
OSLa volume 9(2), 2017 [20] cathrine fabricius-hansen, hans petter helland, anneliese pitz reflexive subsystem is underspecified with respect to possessor properties, the possessor is still unambiguously identified as the subject of the clause.Note, finally, that even closely related languages exhibit differences in their systems (e.g.German vs. Norwegian, English vs. German), while languages from different language families may exhibit important similarities (e.g.Norwegian and Russian with respect to the reflexive/non-reflexive distinction).
[3] o t h e r d i m e n s i o n s o f p o s s e s s i v e c o n t r a s t : s y n ta x , s e m a n t i c s , u s a g e In this section we briefly present some non-morphological dimensions of contrast that are relevant with respect to our object languages but which we do not yet plan to take up in our L2-oriented empirical research.
[3.1] Binding properties Norwegian possessives are not complex for the reasons mentioned in the preceding section alone.They also, in some variants of Norwegian, display binding properties that go well beyond the standard cases mentioned in section [2.1].According to principle A of Binding Theory, reflexive possessives ought to be bound within their binding domain, which normally corresponds to the smallest clause that contains the possessive DP and its co-indexed antecedent; see examples ( 21)-( 22 The reflexive possessive (sitt) of the embedded clause subject in (34) is tested for reflexive binding from the matrix subject (regjeringen).These kinds of sentences are rejected by most speakers in the ScanDiaSyn-survey, and also by the Norwegian author of this paper, but, strikingly, informants from some parts of Norway, especially central parts (Sør-Trøndelag and Nord-Trøndelag) and even some Northern parts, are much more liberal with respect to their acceptability.In general, sentences like (35) containing an irreflexive possessive in the embedded subject position are more readily accepted by the wide majority of speakers: an l2 perspective on possessives: contrasts and their possible consequences [ flertall.majority 'The government do not expect that its proposal (the proposal of the government)/their proposal (the proposal of the members of the government) will get majority.'As stated by Lundquist (2014b, 500), it is however worth pointing out that sentence #156 [our (35)] gets higher scores than sentence #157 [our (34)] in the area where L(ong-)-D(istance)B(inding) in general is quite acceptable (Sør-Trøndelag, Møre og Romsdal, northern Oppland and northern Hedmark).Thus, we have to accept that the already complex distribution pattern of Norwegian reflexive and non-reflexive possessives in normative cases are further complicated by the existence of LDB in varieties of Norwegian.In Julien (2015), we find many cases of binding into embedded clauses, in subject (37) or non-subject position (36), with a surprisingly high rate of acceptability, for example: (36) Hun Since it is her own book Shani Boianiju wants to promote, the reflexive possessive si* is called for while in the next example, the reflexive si* would refer to the authorities (the clausal subject), which of course is not intended: (39) Svenske myndigheter fratok i gar Mijailo Mijailovic i *sitt i (√hans i ) svenske statsborgerskap. [22] cathrine fabricius-hansen, hans petter helland, anneliese pitz Lit.: 'The Swedish authorities yesterday deprived Mijailo Mijailovic (of) his Swedish nationality.'(Aftenposten 2004) Such complexities in the grammar(s) of Norwegian should be borne in mind when we discuss L2 acquisition challenges involving Norwegian as L2 (section [4.6]).
[3.2] Possessives vs. definites: (in)alienable possession, and 'external' possessives There are further highly interesting questions about the distribution of possessives in our languages that we mention en passant without going into details.Thus, as we saw in section [1], 'our' possessives combine definite reference (anaphoric or deictic) with a relational meaning of possession in the broad sense (hierarchy, property, kinship, part-whole, and so on) (Baron et al. 2001;Heine 1997;Koptjevskaja-Tamm 2002;Zifonun 2005).In this respect, the type of relation marked by a possessive DP is typically alienable.My house is a type of (concrete) object that I may own for possibly a very long period, but I may also sell it.Alienable possession is contrasted with inalienable possession, which encodes parts -typically body parts -that are intrinsically linked to the 'possessor'.Inalienable possession is of course not possession in the strict sense.I do not own my arm, leg, nose and so on, but the parts of my body are inherent parts of me.Inalienability is thus related to inclusion and dependency.What is in the part is necessarily in the whole and the whole includes the part (Kleiber 2008).Since all human beings in principle have the same body parts, inalienable possession is typically marked as presupposed or given information.This is why some languages tend to express this type of relation by a definite determiner, cf.(40). 20On the ontology scale of Kleiber (2008), humans have the largest amount of inalienable parts and properties the least: Humans > animals > concrete objects > events > properties.
( an l2 perspective on possessives: contrasts and their possible consequences [23] The head and the eyes are integrated (body) parts of the subject referent.Hence the use of definites in both Norwegian, French and German.However, there are cases where the explicit marking of inalienables differ from language to language: (41) a. Norwegian In a case like (41), one might ask why possessive marking should be necessary in French in contrast to Norwegian and German.Whenever the speaker feels the need to establish the possessive (in the wide sense) referential link with the antecedent in an explicit manner, a possessive may be used.This tendency seems to be stronger for English and French than Norwegian (Woldsnes 2013), 21 and also stronger for German than Norwegian.In fact, more generally, our languages exhibit differences of usage in this area, which we cannot pursue further in the present paper: ( Und sie wollte nach der Mutter rufen, sie nahe bei sich haben.(HW1TD.1.s7)√Und sie wollte nach ihrer Mutter rufen, sie nahe bei sich haben.
We will not be concerned with usage preferences between possessives and def- [21] See also Hasselgård (2012) concerning possessive absolutes in English and Norwegian.
cathrine fabricius-hansen, hans petter helland, anneliese pitz inites in our object languages.We do have to take into consideration, though, that the grammatical possessive system for a language like Norwegian is highly complex in itself, opening up for dialectal and idiolectal variation.This makes it hard to acquire even for native speakers of Norwegian.These points should be borne in mind when investigating the acquisition of e.g.German and French L2 by Norwegian L1-speakers and the acquisition of Norwegian as a foreign language.
Another means to express the inalienable relation between possessor and possessum is the use in some languages of so-called external possessors (Haspelmath 1999;Stolz et al. 2008;Zifonun 2005).In the external possessor construction, the possessor is realized as a constituent of its own, in German as a dative noun phrase called the possessive dative or Pertinenzdativ (Zifonun et al. 1997; cf. ( 43a) and ( 44) from (40c) above.In French, the external possessor appears as a clitic (43b).According to Lødrup (2009a), the external possessor in Norwegian (Icelandic, Swedish and Danish) typically appears as a PP with a locative preposition ( 45 Lit: 'They had to remove the liver on him.' 'They had to remove his liver'. Another construction strongly resembling the dative external possessor in German and French results from possessor raising in Norwegian (Lødrup 2009b) (46) -and English.
(46) Hun slo ham i hodet.'She hit him in the head.' This construction, however, differs from the dative external possessor construction in German by having the possessor as a direct object while in German, the possessor is realized as a dative and the possessum typically as an accusative, i.e. direct, object; and likewise for the French external possessor construction (Lødrup 2009b).
A comparison of external possessors in Norwegian and German fiction in the OMC (see footnote 20) revealed that the German dative possessor occurs more often than the Norwegian external possessor PPs (Holthe 2016).
an l2 perspective on possessives: contrasts and their possible consequences [25] [3.3] Categorial status and definiteness Typically, even in theoretically driven accounts (Alexiadou et al. 2007;Lødrup 2011;Peteghem 2012), possessive 'pronouns' may be viewed as either 'determiners' or 'adjectives'.Their determiner-like behavior is easy to demonstrate.Thus for English, French, German and Norwegian, the prenominal possessive is in complementary distribution with genuine determiners, as witnessed in (47); more specifically, the possessive makes the DP semantically definite like a definite article (Zifonun 2005).Hence the definite D and the (prenominal) possessive D may be said to occupy the same slot in syntactic structure.This, of course, is a more general typological tendency (for cross-linguistic data, see Alexiadou et al. (2007, 566ff)) ruling out their co-occurrence.In general, languages don't accept definites co-occuring with possessives in front of the head noun.At the same time, there are languages that do have definite -or even indefinitedeterminers co-occuring with possessives.Italian is a case in point (Cardinaletti 1998), showing a definite determiner (il) preceding the possessive (suo) and the (expressed) noun: (48) il suo libro The possessive in (48) (suo) exhibits adjectival properties.In a similar vein, French has possessives with adjectival morphology, which are necessarily preceded and followed by respectively a definite determiner and an elliptical head noun: (49) le sien (*livre) In generative grammar, these issues have received much attention from the 1990s and onwards; see, e.g.Julien (2005) and Alexiadou et al. (2007) for an overview.
They have also been framed as a distinction between strong, weak and clitic forms (Cardinaletti 1998).The main idea of this type of approach is that all possessives, whether they are determiner-like or adjectival-like, share the same base position.Assuming a fine-grained elaborate structure of the nominal projection (= DP), the possessive starts out as a specifier of a lower constituent within the extended DP, for example the nP.This corresponds to the strong position of the possessive in Cardinaletti's terms.Adapting her framework slightly, the nominal head will move from the low NP into the head position of the nP and then further up to the head position of a functional projection (FP) above and to the left of the possessive: cathrine fabricius-hansen, hans petter helland, anneliese pitz While the strong possessive, as in (50), remains in its base position, weak possessives move further up the tree, to a higher (functional) specifier position: Besides being preceded by a definite article, weak possessives in Cardinaletti's system are thus deficient.However, they still occupy a specifier position (of FP).This property distinguishes them from fully grammaticalized clitics.The French system is given as an example in Cardinaletti's treatment of a clitic possessive, adjoining to the head position of D as in ( 52): The strong-weak-clitic-treatment of possessives -and more generally of personal pronouns (Cardinaletti & Starke 1999) has been quite influential, but it has mainly been applied to Romance. 22 For our purposes, we will not make use of the strong-weak(-clitic) distinction.It should be noted however that the Norwegian system is particularly complex from a syntactic viewpoint since Norwegian has both prenominal (53a) and postnominal (53b) possessives.In this respect Norwegian differs from Danish, where possessives are restricted to the prenominal position.As for the French and German case (see above), we treat the prenominal possessive as a kind of determiner, or at least as occupying a head position, possibly of a possessor phrase, high in the extended nominal projection.The postnominal possessive of Norwegian however behaves quite differently since it combines obligatorily with a definite noun: (53c) is ungrammatical.
(53) c. *bil car hans his [22] Attempts have been made, though, to adapt it to Germanic.Lødrup (2011), for instance, takes Norwegian postnominal possessives to be weak and Norwegian prenominal possessives to be strong, contrary what one would expect in view of Cardinaletti's hypothesis for Romance.In Norwegian, only prenominal possessives can be coordinated (mitt og hennes hus 'my and her house') and focused (dette er bare MITT hus 'This is only MY house').And even if postnominal possessives may also easily be focused (bilen MIN, ikke DIN 'my car, not yours'), this is not taken by Lødrup (2011) as an argument for necessarily treating them as strong.
an l2 perspective on possessives: contrasts and their possible consequences [27] This means that the postnominal possessive in Norwegian cannot adjoin to the D-position like the French possessive determiner in Cardinaletti's treatment.It should rather be seen as a specifier of a lower nominal projection, either the nP (Cardinaletti 1998) or the NP (Julien 2005).
[4] l 2 a c q u i s i t i o n c h a l l e n g e s i n t h e c o r e s y s t e m s : g e n e r a l a ss u m p t i o n s [4.1] Preliminaries As mentioned in section [1], we assume that (i) achieving native-like fluency in the use and processing of L2 possessives (in the languages we are concerned with) is a task of varying complexity, depending in part on the degree of isomorphism between the possessive core systems of the specific L1/L2 pair involved; and (ii) for a given L1/L2 pair, some possessive contrasts may be more fundamental or pervasive than others, impeding L2 performance of even quite advanced L2 learners.
It should be stressed, though, that production and comprehension (interpretation) are very different tasks (Zeevat 2014) and, accordingly, that features of L2 which cause problems in L2 production may represent less of an obstacle in interpretation tasks, and vice versa (see e.g.Jarvis & Pavlenko (2008, 15ff)).Also, when measuring L2 proficiency, it is important to distinguish between the outcome of a production or interpretation task, i.e. the product, on the one hand and what we shall call the process, i.e. processing itself, on the other hand. 23In our context, this means that ideally, hypotheses concerning the acquisition or command of L2 possessives should be specified along these different dimensions and tested in adequate offline and online experimental settings.
In addition to free production and interpretation of written or spoken L2, foreign language learners are often faced with the task of explicitly translating between L1 and L2 (both ways).Accordingly, we consider (product and process) data from translation tasks to be relevant in our context as well, despite the special character of such tasks.Notably, translation involves either L1 comprehension and L2 production or L2 comprehension and L1 production, depending on whether L2 is the target or the source language.Consequently, one might expect form-based priming between the two languages to be more frequent in translation tasks than in non-translational L2 production or comprehension; cf.Pitz et al. (2017), Helland (2017) and Behrens (2017). 24  [23]  We prefer the term process since processing, as used psycholinguistic literature, seems to used partly in a very general sense but also more specifically for the 'decoding' (comprehension) process.[24] Relevant German literature makes a convenient terminological distinction between Hin-Übersetzung (L1 to L2) and Her-Übersetzung (L2 to L1).
In the sections [4.2] through [4.5] we summarize what we take to be the main possessive challenges of English, German, French and Russian as L2; section [4.6] is concerned with Norwegian as L2 (first column of table 3), and in particular with the possible effects of the distinction between reflexive and irreflexive possessives.
At present the research questions we pursue in ongoing and planned empirical investigations focus on the boldface pairs in table 3, which have Norwegian as L2 or Norwegian as L1 with German or French as L2, in part for practical reasons (availability of data and test persons), in part because these three languages are underrepresented in L2 research, as compared to English.In addition, experimental research concerning Czech vs. German and Polish vs. English possessives is being conducted in Dortmund and Warsaw, respectively.
[4.2] English as L2 The fairly simple English possessive systems may be assumed to represent less of a L2 challenge to Norwegian, German, French and Russian learners than the other way round.Note however that genuine restructuring (Jarvis & Pavlenko 2008;McLaughlin 1990;Pashler 1999) is demanded from French learners of English: In the possessor singular they will have to replace their possessum-oriented gendernumber distinction (son/sa/ses, see figure 2) by the possessor-oriented distinction between his, her and its, which is of a semantic nature in the sense that it reflects properties of the possessor referent (± human, natural gender); cf.figure 1. 25 That is, viewed from the L2 production perspective of a French En2 learner, each of the three L1 items son, sa and ses may correspond to either his or her (or its) in the [25] Note, though, that French does exhibit possessor-oriented stem variation with respect to the categories person (first, second, third) and number (cf.table 1).
Since Norwegian has the possessor gender and ±human distinction built into its system, albeit at the lower hierarchical level of irreflexive possessives (cf.hans 'his', hennes, 'her' and dens/dets 'its') (figure 4), Norwegian learners of English, on the other hand, may be expected to acquire the English possessive system quite easily.This holds for German learners, too, due to the fact the possessives sein* and ihr*, which normally reflect the grammatical gender of the antecedent DP (if singular), may be used deictically referring to a male, female (or human plural) possessor, respectively; similarly for the Russian irreflexive possessives ego, eë and ich.
[4.3] German as L2 As far as German possessives are concerned, their complicated inflection, based on number, gender and case agreement with the head noun, is a well-known problem for Ge2 learners, more or less independently of their L1 -and for Ge1 acquisition as well.Being common to German determiners in general (apart from some details), however, it is of less interest in our context than two possessorrelated characteristics of the German possessive system: the fact that the choice between sein* and ihr* is determined by number and grammatical rather than natural gender of the antecedent DP; and the fact that ihr* is underdetermined between (possessor) plural and femine singular (see section [2.4] and figure 3). 26 Thus Norwegian and English Ge2 learners will have to substitute a possessorrelated distinction of grammatical gender for their native semantic distinction and in addition adapt to an idiosyncratic possessor-related ambiguity.The latter, notably, may be more of a problem in Ge2 interpretation than in Ge2 production: No1 hennes (irreflexive possessor singular feminine), deres (irreflexive possessor plural) -and (reflexive) si* with a singular feminine or plural possessor -'converge' to Ge2 ihr* (production perspective) while Ge2 ihr* may correspond to No1 si* (reflexive), hennes (irreflexive singular feminine) or deres (irreflexive, plural) (comprehension perspective).
French Ge2 learners face a similar restructuring challenge as with En2 (see above), possibly complicated by the phonological similarity between German sein (possessor singular masculine) and French son (possessor singular, unspecified possessor gender; possessum singular masculine).Consequently one might expect French Ge2 learners to use sein* instead of correct ihr* more often than the other [26] Ordinary third person pronouns distinguish between masc.er* and neuter es in the singular while sie* like the possessive ihr* may take a sg.fem.or a plural antecedent.Note also that the word form ihr in addition to (possessum) sg.masc.nominative and sg.neut.nominative/accusative of the possessive ihr* represents the dative of the third person fem.sg.pronoun sie* 'she' and the nominative of the second person plural pronoun, i.e. 'you' (plural).
cathrine fabricius-hansen, hans petter helland, anneliese pitz way round.In a similar vein, Norwegian and Like Norwegian Ge2 learners, Russian Ge2 learners -even when aware of the systematic differences -may tend to associate sein* with their native reflexive s-possessive (si* and svoj*, respectively), resulting in a skewed distribution of error types in Ge2 production (sein* used instead of correct locally bound ihr* more often than vice versa) and Ge2 comprehension (sein* more often than ihr* erroneously understood as locally bound, i.e. interpreted reflexively); for No1/Ge2 see Pitz et al. (2017).
[4.4] French as L2 Turning to L2 French, we assume that English, Norwegian and Russian learners may tend to erroneously equate the possessor-dependent gender-number distinction between their possessives -English his, her, (its,) their, Norwegian irreflexives hans, hennes, (dens/dets) and deres, and Russian irreflexives ego, eë and ich -with the possessum-related gender-number distinction between son, sa and ses, ignoring the French possessor-plural possessive leur(s).
Likewise, German Fr2 learners will have to replace their possessor-dependent gender-number distinction (sein*/ihr*) by the French possessum-dependent gender-number distinction (son/sa/ses); and due to the greater phonological similarity, we might expect that they wrongly equate son/ses with sein* (ignoring that e.g.son père is ambiguous between sein Vater and ihr Vater, as is ses enfants between seine Kinder and ihre Kinder).On the other hand, since sa corresponds to the sg.fem.form la of the definite article, Fr2 learners may tend to equate sa with the feminine possessor ihr*, ignoring the ambiguity of sa mère between ihre Mutterseine Mutter.Like German Fr2 learners, Norwegian and Russian Fr2 learners will have to 'deactivate' any associative bond between the French s-possessive and their native reflexive s-possessive (si* and svoj*, respectively); for No1/Fr2 see Helland (2017).
[4.5] Russian as L2 As for Russian as L2, we assume -abstracting from inflectional complicationsthat English, German and French Ru2 learners encounter the same kind of problems with the reflexive possessive (svoj*) vs. irreflexive third person possessives (ego, eë, ich) as they may be expected to have when learning Norwegian (see section [4.6]) while Norwegian Ru2 learners should come to master the choice between svoj* and an irreflexive third person possessive -and the obligatory reflexive (locally bound) interpretation of svoj* -quite easily.
[4.6] Norwegian as L2: reflexivity and (potentially) false possessive friends Abstracting from the inflectional complexities of German and Russian, the Norwegian core system is evidently more complex than any of the other possessive OSLa volume 9(2), 2017 an l2 perspective on possessives: contrasts and their possible consequences [31] systems presented in section [2], primarily due to the distinction between reflexive and irreflexive possessives but also because singular irreflexive possessives are differentiated according to both natural and grammatical gender of the possessor (figure 4). 27Acquiring Norwegian possessives, then, would seem to demand more complicated restructuring from L1 speakers of German, French or English than the other way round since the reflexivity parameter introduces a type of possessorrelated distinction (possessor cue) at the highest 'decision' level which is absent in their L1 (compare figures 1, 2, and 3 to figure 4). 28 The result is lexical L1-L2 divergence involving Norwegian si*: In production, No2 learners have to choose between si* and some irreflexive alternative (hans/hennes/…), depending on the syntactic (binding) circumstances, irrespective of which L1 possessive -if any (see section [3.2]) -would be adequate in the given context.As for comprehension, German and French No2 learners conversely face the challenge that the Norwegian si* variants (sin, si, sitt, sine) are also underdetermined in the sense of being unmarked for possessor-related features that determine the choice between lexical alternatives in their L1, i.e. sein* vs. ihr* in German and the s-vs.the leur-possessive in French.Thus the Norwegian, German and French possessives beginning with s--and in particular the specific forms sin (Norwegian), sein (German) and son (French) -constitute what may be termed (partly or potentially) 'false friends': they are morpho-phonologically similar but cued differently in relation to the possessor.
We expect the No2 performance of German and French learners to reflect the hierarchical importance of the reflexivity contrast in general and the false-friend relation between No2 reflexive possessives and L1 s-possessives in particular.More specifically, we assume that not too advanced No2 learners tend to prefer Norwegian si* over irreflexive alternatives under conditions calling for an s-possessive in their mother tongue and to interpret Norwegian si* in accordance with its (partly false) possessive friend in their L1, whether or not this a correct solution to the production or comprehension task at hand.Specifically, this seems to be a plausible hypothesis for translation tasks (into or from No2), where the actual presence of an s-item in the L1 or No2 source text may prime for the possessive (but potentially false) friend in the target language.And we assume that the possessive performance of otherwise quite advanced No2 learners tends to be somewhat impeded in related ways as compared to the performance of No1 speakers.
As for Russian No2 learners, on the other hand, we do not expect them to have particular problems with the reflexivity distinction in Norwegian while the twofold differentiation of singular irreflexive possessives (hans vs. hennes and dens vs. dets) [27] An additional complexity is the prevailing postnominal position of Norwegian possessives; cf.Anderssen & Westergaard (To appear).[28] Note though that the reflexive-irreflexive distinction is present in the non-possessive pronoun systems of these languages (cf.German sich, French se, English himself /…). [32] cathrine fabricius-hansen, hans petter helland, anneliese pitz may represent a challenge here (compare figure 5 and 4) -and perhaps for French and German No2 learners as well.
[5] s u m m a r y a n d o u t l o o k The present paper set out to present the contrastive background and the basic objectives of a cross-linguistic research project (POSS) that takes an L2-oriented perspective on third person possessives in English, Norwegian, German, French and selected Slavic languages, focusing on L1/L2 pairs involving Norwegian as L2 or L1 but with a view also to pairs including Czech or Polish.
In the first, descriptive part of the paper we first compared the various morphosyntactic ('core') systems of possessives, Russian representing the Slavic group (section [2]), and then briefly presented other -purely syntactic, semantic, or usage-oriented -dimensions of contrasts that go beyond the immediate scope of our project (section [3]).
Section [4], finally, addressed our main research question: What impact -if any -may the contrasts laid bare in section [2] have on the acquisition, processing and use of L2 possessives across our object languages?To what extent and in what ways may the acquisition or command of possessives in L2 be impeded or enhanced by the specific properties of the possessive system in L1?
Our basic assumption is that the challenges a L2 learner faces in relation to L2 possessives depend in part on the degree of isomorphism between the possessive core systems of her/his L1 and the L2 in question.Specifically, the reflexiveirreflexive bifurcation of Norwegian (and Slavic) third person possessives is based on a purely syntactic distinction (local versus non-local binding by the possessor) at the top level of the dimensions determining the choice of third person possessive in production and the search for an antecedent in comprehension.Acquiring this system, then, demands high-level possessive restructuring by German, French and English learners, whose L1 lacks that distinction in the possessive systemalthough reflexivity is explicitly encoded in the system of 'ordinary' third person pronouns (cf.German reflexive sich, French se, English her-/himself etc.).On the other hand, English, German, Norwegian (and Slavic) learners of French have to adapt to the fact that possessor gender, an important dimension in their L1 possessive systems, is irrelevant in the French system (cf.figure 3 and section [4.3]).
We assume that even when L2 learners are fully aware of such important contrasts, i.e. have successfully restructured their explicit knowledge of possessives, their actual L2 performance may still be somewhat hampered as compared to native speakers of the language in question.That is, we assume that automatization may lag behind even for quite advanced L2 learners, in particular -or at least in particular ways -where false possessive friends are involved (e.g.Norwegian si* versus German sein* and French s* (son/sa/ses), German sein* versus French s*).
(ii) Type 2: Comparisons of the use of a particular feature in the IL2, the L1 and the L2.
(iii) Type 3: Comparisons of the use of a particular feature in the IL2 of learners from two or more different L1 backgrounds.
(iv) Type 4: Comparisons of the use of a particular feature in the IL of learners who have two L1s (i.e. are bilingual).
(v) Type 5: Two-way comparisons involving learners with different L1s, each learning the other's L1 as an L2.
Our research evidently relates to type 1, type 2 as well as type 5 in Ellis' classification The notion of transfer, its conditions and manifestations, is an object of ongoing debate, which has been made even more complicated during the later years by (more and more) seriously taking into account that L2 acquisition and fluency -the development of an IL2 -may be influenced not only by L1 but also by other languages learnt before or along with the L2 in question (De Angelis 2007; Westergaard et al. 2016).This is a complication we have neglected here (cf.footnote 5).Even so, though, we take it that empirical investigations confirming or refuting our assumptions will help understand whether and how the relationship between comparable subsystems of L1 and L2 may influence L2 acquisition by speakers of L1.Specifically, since the possessive systems differ across 'our' languages in more interesting ways than the systems of ordinary pronouns and since these two systems differ with respect to the reflexivity parameter within some but not all of 'our' languages, further research along the lines suggested above may shed new light on the division of labor between L1 influence (transfer) and general learner effects in L2 acquisition.
[34] cathrine fabricius-hansen, hans petter helland, anneliese pitz To that end, however, our assumptions must be spelled out in specific, testable hypotheses along the lines indicated in section [4.1], i.e. differentiating between production (free production and translating from L1 into L2) and comprehension/interpretation (including translating from L2 into L1) on the one hand, and between product and process(ing) on the other hand.Learner language studies of the more traditional kind -e.g.studies of L2 learners' use of possessives in written L2 essays and translations into L2 -belong to the production-product category while investigations of how L2 learners understand possessives occurring in L2 texts are of the interpretation-product type.Process-oriented hypotheses in their turn, whether relating to production or comprehension, call for online experiments involving or eliciting the use of possessives; see e.g.Felser & Cunnings (2012); Patterson et al. (2014); Roberts et al. (2008); Schimke et al. (2015) and further references therein for experimental studies on so-called anaphoric resolution of non-possessive anaphors (third person pronouns).In addition, testing L2 learners' explicit grammatical knowledge of L2 possessive will be needed.Relevant preliminary investigations relating to German vs. Norwegian and French vs. Norwegian are presented in Pitz et al. (2017) and Helland (2017) while Behrens (2017) an l2 perspective on possessives: contrasts and their possible consequencesC A T H R I N E F A B R I C I U S -H A N S E N , H A N S P E T T E R H E L L A N D , A N N E L I E S E P I T Z University of Oslo a b s t r a c t figure 2: The French system of third person adnominal possessives.
figure 5: The Russian system of adnominal possessives with third person possessor.
his book b. *le son livre c. *das sein Buch d. *den hans bok looks at translation from En2 into No1.r e f e r e n c e s
Dmitri and Mary love Peter's and Anna's dog.'In contrast to Norwegian, the Russian reflexive possessive svoj* is not restricted to third person subjects but may take first and second person antecedents as well, competing with the regular Indo-European first and second person possessives moj* 'my', tvoj* 'your sing ', etc. (see table 1); cf.(32) and (33).