Variation in the left periphery and complex clauses V2 constraint.
While adverbial phrases in sentence initial position (with the exception of da, ‘there’) are attached to the available SVX format in file 1 (cf. (499) above), their integration into the target V2 format occurs already in file 2 (see example (511)) (note, though, that the verb fails to correctly agree with the subject, the only exception to our previous observation about Maria’s target-like verb inflection with conjoined subjects). We remarked previously on the variation of XVS, VS, XVX and V3 formats in Maria’s written German narratives. File 2 contains no VS orders, but one XVX pattern (compare (513) in which the subject is dropped). As for V3, there is a total of 3 constructions vis-a-vis 7 instances of XVS sentential formats. The proportion of XVS formats thus clearly exceeds that of target-deviant V3 patterns as it occurs in all other files, with the exception of file 4 (compare also Figure 4.10 above). Nevertheless, the alternation of target V2 and target-deviant V3 in files 2-4 (see examples (511)-(512) produced in file 2 and examples (514)-(515) produced in file 4) suggests that the availability of nonsubject V2 does not go along with the immediate exclusion of alternative formats. This conclusion is in line with what is known about the attainment of V2 by other L2 learners of German. It is interesting to note in this context that main clause V3 patterns appear in files 3 and 4 only with the adverbials plotzlich (‘suddenly’) and dann (‘then’), two elements that typically serve the function of connecting narrative episodes; at the same time, we acknowledge that these elements also appear in target-like XVS patterns.
Incidentally, example (512) above shows that the selective properties of verbs taking infinitive clausal arguments are not yet mastered (alternatively, the verb wunschen, ‘wish’, might appear with a finite verb, but in the context of an embedded clause introduced by dass, ‘that’).
Verb inflection. Turning to verb inflection in file 2 the analysis reveals that it continues to be characterised by the alternation of target-like and target-deviant forms. The sequence in example (516), in which the copula verb ist (‘is’) appears in combination with the main verb bleibt (‘stays’) might reflect a lack of attention during the production process; alternatively, it might represent an error in the past participle formation of the verb bleiben (‘stay’), that is, geblieben (‘stayed’) (perfect tense with this verb involves the auxiliary sein). However, if we look at the verb forms produced in example (517), where, by assumption, Maria also intended to produce a perfect tense form, we might speculate that Maria still lacks the knowledge of the constraints on perfect tense formation with the auxiliary verb sein (‘be’). Whereas the target perfect tense form would require the combination of the auxiliary ist with a past participle form of the main verb, a finite main verb form appears in combination with ist in example (516), and in a non-finite (infinitive) form (517). Against this backdrop, we conclude that although Maria’s learner grammar is quite advanced with regard to structure-building, the domain of inflectional morphology is subject to variation.
Finally, example (518) is a remarkable infinitive construction; however, the choice of the wrong modal verb form will (‘wants’) suggests the inflection paradigm of this modal verb is not mastered. Example (496) shows that constructions with some separable verbs are not yet fully mastered as the separable prefix appears twice in this construction.
Interrogation. Turning to question formation at this stage, the sequence in (520) is an example of target-like question formation, with the target-like choice of the 2nd person singular expletive form of the copula sein.
Embedded clauses. Maria produces embedded clauses with the complementiser dass (‘that’) and a target-like sentence-final verb placement in file 3 (cf. (521)). However, given that there is no further instance of such complex structures in files 4 and 5 we can only speculate on the implementation of the target-like head- final IP. Nevertheless, structures like (521) are remarkable as they do not occur in the narratives of the other participants. Neither do constructions with impersonal or expletive pronouns. Again, Maria’s production of a construction with the impersonal pronoun es (‘it’) in file 3 (cf. (522)) marks an exception.
In this context a note is due on another phenomenon that reveals Maria’s advanced knowledge of German. Indeed, the analysis of the file 3 narrative also reveals a sophisticated use of prepositions (compare examples (523)-(526)), to indicate locations and the relation between the verb and the complement. Although not all constructions are target-like they contrast markedly with the use of auf (‘on’) observed in other written German productions of this corpus. In fact, in Maria’s narratives we do not find evidence for the use of auf as a case marker with verbs that do not subcategorise for it in target German.
Implementation of V2
Eventually, in file 5, constructions with non-subjects in sentence-initial position adhere to the V2 constraint across the board. Sequences like (527a-c) show that target-deviant subject drop continues to be produced until file 5, which suggests that the correct setting of the pro-drop parameter remains a task to be tackled. Finally, the target-like yes-no question in (528) is illustrative of the availability of the mechanisms for question formation (including verb raising and a structural layer above the IP, i.e. the CP).