Subordination and question formation
The expansion of the available structural format by the projection of the CP layer is commonly tied to the production of embedded clauses introduced by a complementer and target-like question formation. Consequently, in our analysis of the data we looked at the incidence of complex sentential structures and question formation. Now both phenomena, as became apparent in the discussion of the participants’ developmental profiles, occur fairly infrequently in the data. Recall that for some participants we concluded that there was no sufficient evidence to conclusively establish the status of subordination and question formation in their learner grammars.
Subordination. Turning to subordinated clauses, it is interesting to note that the complementiser weil (‘because’) is produced early on. Typically, it appears in combination with verbless clauses at the time when the IP is not yet available, as is the case in Hamida’s file 1 or Fuad’s file 1. By assumption, at this stage, weil is adjoined to the available VP structure (recall our previous comments regarding the use of functional items despite the lack of the associated target grammatical properties). Upon the availability of the IP, word order in weil-introduced clauses mirrors main clause word order, which suggests that the CP is projected on the basis of the available head-initial IP. Note that in target German weil is the only complementiser that allows for main clause word order, but this option is restricted to SVO order. Evidence of target-like sentence final verb placement in embedded clauses is rare in the narratives collected as is the use of complemen- tisers other than weil. Some learners use wh-word introduced embedded clauses in which the order is the same as in the equivalent direct questions or involves the drop of the auxiliary as is typical of Hamida’s productions at the time. Fuad produces one dass-introduced embedded clause in file 4, but word order suggests that the IP is head-initial in this case, too.
In summary, only Maria appears to have implemented the full CP structure in her learner grammar: non-subject V2, question formation and target-like embedded clauses are productive in her data. One learner, Simon, does not produce any evidence of the availability of these grammatical processes. For all other learners we can only speculate on the availability of a CP layer: weil remains the only complementiser used productively and question formation is restricted to the pattern “wh-word + ist”. If weil-introduced clauses involve a CP we assume this is added to the head-initial IP available. This sentence structure not only resembles that of symmetric V2 languages like Yiddish and Icelandic (cf. Vikner 1995), it is also attested in the learner grammars of adult learners of L2 German (cf. example (699)) (Plaza-Pust 2000: 244f.). Recall in addition (cf. section 4.3.3) that individual variation has also been found to occur occasionally in the productions of children acquiring L1 German. Embedded clauses with complementisers other than weil exhibit target-like word order in Christa and Fuad. Yet the few instances produced are insufficient to establish whether the IP in CP structures is set to the head-final value in their learner grammars.
Question formation. The lack of the mechanisms necessary for target-like question formation is reflected in the predominance of formulaic questions such as wo ist (‘where is’) and wer ist (‘who is’). Christa and Fuad produce questions with the second person suppletive form of the copula verb sein. As the question is the same as the title of the story and no other instances are produced we can only speculate on whether the necessary mechanisms are productive. It seems plausible to assume, however, that they are “within reach”.^ Only Maria produces yes-no questions which provides further evidence that the mechanisms necessary for question formation are productive (compare example (700) repeated here for convenience).
15 This holds especially in the case of Fuad who produces the first instances of non-subject V2 in file 5. Note, however, that Christa does not produce such sequences in the narratives collected.
The observation that the type of structures borrowed basically reduces to DGS- like idiomatic expressions once the IP is established is indicative of the circumstance that borrowing at the structural level is not required at this stage. The only mixed grammatical property that continues to prevail in the stories at a more advanced level concerns the overgeneralisation of auf, a phenomenon that comes as no surprise given the continuing lack of the target agreement and case marking paradigms. Further, the use of verbless clauses at this stage shows (a) that previous, more elementary grammars continue to be available and (b) that lexical gaps are filled by borrowing expressions from DGS.
-  For proponents of the symmetric structure of German, the production of non-subject initialV2 clauses also involves the availability of a CP.