Abstract
Many models of the nonnative acquisition of syntax build on the assumption that target lexical projections are developmentally prior to target functional projections (e.g. Minimal Trees (Vainikka & Young-Scholten 1994), Modulated Structure Building (Hawkins 2001)), or that learners only have to grapple with the acquisition of the topmost levels of syntactic structure (e.g. Vulnerable C-domain (Platzack 2001)).
Germanic verb second (V2) is often said to be difficult for L2 learners irrespective of their L1. Targetlike finite verb placement in non-subject-initial root clauses (V2) has been described as being dependent on targetlike nonfinite verb placement (VO/OV) having been acquired first. Italian, Spanish, Portuguese, English, Swedish as well as Turkish and Korean learners of German have been reported to acquire German-style OV long before V2 (e.g. duPlessis et al. 1987; Vainikka & Young-Scholten 1996b; Pienemann 1998; Schwartz & Sprouse 1994; Vainikka & Young-Scholten 1994, 1996a, b).
Recent work on Swedish/German (e.g. Vulnerable C-domain (Platzack 2001), Processability (e.g. Håkansson, Pienemann & Sayehli 2002)) claims that speakers of a V2-language learning another V2-language start with a non-V2 interlanguage grammar, i.e. they do not transfer V2, but follow a universal developmental path of verb placement.
The present paper contests the above claims, presenting quantified oral production data from 6 L1 Swedish adult ab-initio learners of German (4 & 9 months of exposure to classroom German) as well as oral and written production data from 23 L1 Swedish 16-year-old intermediate learners (3 years of classroom German). I document productive use of non-subject-initial V2 declarative clauses after only 4 months of exposure to German, at a time when the informants’ interlanguage syntax elsewhere is nontargetlike – and this, notably, includes head-initial VPs (VO).
There is also a categorical difference between informants depending on whether they have prior knowledge of L2 English. Of the Swedish adult ab initio learners, 3 were acquiring German as their first L2, and they produce 100% targetlike V2. However, the other informants who had substantial previous exposure to English, only produce 45% targetlike non-subject-initial V2 in their L3 German, also allowing nontargetlike V3, i.e. Adv-SVX (cf. Bohnacker 2005a, 2005b). This suggests that the non-V2 syntax of their L2 English influences their acquisition of L3 German, a potential confound that was not controlled for in earlier studies such as Håkansson, Pienemann & Sayehli (2002) and Platzack (2001).
The results suggest that there is no universal developmental route to L2/L3 German verb placement, that the C-domain is not vulnerable per se, that the V-domain is not invulnerable per se, that learners make use of their V2-L1 syntax (Swedish), and that knowledge of a non-V2 language (English) can make it initially more difficult to acquire another V2 language (German). The findings are interpreted as strong empirical support for transfer approaches to the nonnative acquisition of syntax (e.g. Schwartz & Sprouse 1994, 1996).
Going beyond the findings reported in Bohnacker (2005a, b), I look at nonfinite verb placement (VO/OV) in the interlanguage German of the Swedish learners in more detail. Here, the learners show awareness of German being OV early on (by 9 months), in contrast to Swedish.
However, alongside targetlike OV productions, the ab initio learners continue to produce 30% nontarget VO utterances. This also holds for a large subgroup of the 16-year-old intermediate learners who exhibit sizeable percentages of nontarget head-initial VPs (VO) after 3 years of exposure. Yet there is no correlation between VO and V2 violations, nor between targetlike OV and perfect V2. Nonfinite verb placement (VP headedness), especially in oral production, is a lot less targetlike than their finite verb placement (i.e. V2). However, VP headedness appears to interact with syntactic context in quirky construction-specific ways. For instance, when embedding a VP under a finite auxiliary, learners overwhelmingly stick to head-final VPs (cf. (1)), but when coordinating VPs, they tend to produce a head-initial VP in the second conjunct (cf. (2)). I will address some of the implications these findings may have for current models of L2 syntax acquisition.
(1)ich will ein Hund haben.
I want a dog have
‘I want to have a dog.’
(2)L2er: ich will nach Paris fahren und schreiben Poesie und Dichte.
I want to Paris go-INF and write-INF poetry and poems
Target: ich will nach Paris fahren und Poesie und Gedichte schreiben.
‘I want to go to Paris and write poetry (and poems).’
Germanic verb second (V2) is often said to be difficult for L2 learners irrespective of their L1. Targetlike finite verb placement in non-subject-initial root clauses (V2) has been described as being dependent on targetlike nonfinite verb placement (VO/OV) having been acquired first. Italian, Spanish, Portuguese, English, Swedish as well as Turkish and Korean learners of German have been reported to acquire German-style OV long before V2 (e.g. duPlessis et al. 1987; Vainikka & Young-Scholten 1996b; Pienemann 1998; Schwartz & Sprouse 1994; Vainikka & Young-Scholten 1994, 1996a, b).
Recent work on Swedish/German (e.g. Vulnerable C-domain (Platzack 2001), Processability (e.g. Håkansson, Pienemann & Sayehli 2002)) claims that speakers of a V2-language learning another V2-language start with a non-V2 interlanguage grammar, i.e. they do not transfer V2, but follow a universal developmental path of verb placement.
The present paper contests the above claims, presenting quantified oral production data from 6 L1 Swedish adult ab-initio learners of German (4 & 9 months of exposure to classroom German) as well as oral and written production data from 23 L1 Swedish 16-year-old intermediate learners (3 years of classroom German). I document productive use of non-subject-initial V2 declarative clauses after only 4 months of exposure to German, at a time when the informants’ interlanguage syntax elsewhere is nontargetlike – and this, notably, includes head-initial VPs (VO).
There is also a categorical difference between informants depending on whether they have prior knowledge of L2 English. Of the Swedish adult ab initio learners, 3 were acquiring German as their first L2, and they produce 100% targetlike V2. However, the other informants who had substantial previous exposure to English, only produce 45% targetlike non-subject-initial V2 in their L3 German, also allowing nontargetlike V3, i.e. Adv-SVX (cf. Bohnacker 2005a, 2005b). This suggests that the non-V2 syntax of their L2 English influences their acquisition of L3 German, a potential confound that was not controlled for in earlier studies such as Håkansson, Pienemann & Sayehli (2002) and Platzack (2001).
The results suggest that there is no universal developmental route to L2/L3 German verb placement, that the C-domain is not vulnerable per se, that the V-domain is not invulnerable per se, that learners make use of their V2-L1 syntax (Swedish), and that knowledge of a non-V2 language (English) can make it initially more difficult to acquire another V2 language (German). The findings are interpreted as strong empirical support for transfer approaches to the nonnative acquisition of syntax (e.g. Schwartz & Sprouse 1994, 1996).
Going beyond the findings reported in Bohnacker (2005a, b), I look at nonfinite verb placement (VO/OV) in the interlanguage German of the Swedish learners in more detail. Here, the learners show awareness of German being OV early on (by 9 months), in contrast to Swedish.
However, alongside targetlike OV productions, the ab initio learners continue to produce 30% nontarget VO utterances. This also holds for a large subgroup of the 16-year-old intermediate learners who exhibit sizeable percentages of nontarget head-initial VPs (VO) after 3 years of exposure. Yet there is no correlation between VO and V2 violations, nor between targetlike OV and perfect V2. Nonfinite verb placement (VP headedness), especially in oral production, is a lot less targetlike than their finite verb placement (i.e. V2). However, VP headedness appears to interact with syntactic context in quirky construction-specific ways. For instance, when embedding a VP under a finite auxiliary, learners overwhelmingly stick to head-final VPs (cf. (1)), but when coordinating VPs, they tend to produce a head-initial VP in the second conjunct (cf. (2)). I will address some of the implications these findings may have for current models of L2 syntax acquisition.
(1)ich will ein Hund haben.
I want a dog have
‘I want to have a dog.’
(2)L2er: ich will nach Paris fahren und schreiben Poesie und Dichte.
I want to Paris go-INF and write-INF poetry and poems
Target: ich will nach Paris fahren und Poesie und Gedichte schreiben.
‘I want to go to Paris and write poetry (and poems).’
Original language | English |
---|---|
Title of host publication | Language acquisition and development: Proceedings of GALA 2005 |
Publisher | Cambridge Scholars Publishing |
Pages | 65-78 |
ISBN (Print) | 1-84718-028-0 |
Publication status | Published - 2006 |
Bibliographical note
The information about affiliations in this record was updated in December 2015.The record was previously connected to the following departments: Swedish (015011001)
Subject classification (UKÄ)
- Languages and Literature