Modal verbs and their relation to finiteness in Dutch first language acquisition

In Dutch child language, as in other Germanic languages, the first verbal elements are non-finite: lexical verbs used as infinitives or participles (Wijnen et al. 2001, Gillis 2003). During this stage, auxiliaries and modals are rarely found. Starting around the age of two, the first finite elements appear: next to some frequent lexical verbs, the most frequent finite verbs are copulas, (temporal) auxiliaries and modals (Van Kampen & Wijnen 2000, Jordens 1990). In the literature, these first finite verbs are often treated as one category. In the next stage lexical verbs occur both finitely and non-finitely. These nonfinite verb forms have been assumed to have a modal interpretation (see Ingram & Thompston 1996, Hoekstra & Hyams 1998, Blom 2008). Later on, these lexical infinitives are replaced by modal auxiliaries with an infinitival complement (Blom & Wijnen 2000). How this change takes place and why there is a stage in which lexical with having a modal meaning occur next to modal main verbs still remains open. When studying the acquisition of finiteness both morphological and structural characteristics, like word order, periphrastic verb constructions and subject presence need to be taken into account. Dutch modal verbs differ from other verbs on both aspects. Not only is finiteness on modals marked differently (with ablaut and without the typical finiteness suffix in the 3th ps. sg.), their syntactic status is complex as well: Dutch modal verbs can be used both as auxiliary and as main verb. So, from the perspective of finiteness, modal verbs share structural characteristics with (temporal) auxiliaries and with lexical verbs. In the acquisition of modal verbs, two things are remarkable. First, modal verbs are relatively frequent in child language (Stephany 1986). Second, modal verbs appear immediately as finite verbs, in opposition to lexical verbs that first appear non-finitely (Van Kampen & Wijnen 2000). With this as starting point, we examined how modal verbs are acquired, and how this differs from other verb classes and which role modal verbs play in the acquisition of finiteness. In order to investigate this, we analyzed spontaneous speech data from 6 monolingual Dutch children (age 1;10 till 3;6, about 70.000 utterances in total) from CHILDES (MacWhinney 2000). We counted modal verbs and occurrences of zijn, hebben and gaan. Note that zijn can be used as copula or auxiliary, and hebben and gaan, as lexical verb or auxiliary. However, in spontaneous data distinguishing between those categories is impossible. We analyzed whether these verbs were used finitely or not and whether they were used as main verb or with an infinitive or participle. From an average MLU of 2.0, more than 10% of the child utterances contain a modal verb (fig. 1), almost all of them are realized finitely. Modal verbs occur early in combination with an infinitive. Still, the modal verb is more often used as a main verb, which indicates the importance of this category (fig 1). Fig. 2 shows that the moment of appearance and the relative frequency of different modal verbs vary, which suggests that referring to modal verbs as one category passes over some important differences. In contrast with results on German and English (Shatz and Wilcox 1991, Adamzik 1985) we find a high amount of moeten (typically expressing obligation) and a low amount of willen (mostly used for expressing the own will). Comparing modal verbs to zijn, hebben and gaan (fig. 3), we see that zijn (‘to be’), is hardly ever used with an infinitive/participle, and probably often as a copula, and gaan (‘going to’) appears often and early in periphrastic constructions. Modals seem to be in between zijn, as the verb used alone and gaan as most ‘auxiliary’-like verb. Our results show that neither a classification in non-finite and finite verbs and periphrastic constructions, nor grouping modals with temporal auxiliaries and copula (see e.g, Wijnen, 1997, Blom, 2008), does do justice to the special position of modal verbs in Dutch acquisition. We argue that the acquisition of modals in Dutch differs from both lexical verbs and auxiliaries. Moreover, it seems that acquiring Dutch modal verbs means acquiring two different grammatical structures, which both appear from early on. Finally, not only the different grammatical status of modal verbs has to be acknowledged, we also have to distinguish within the category of modal verbs.