9 Functional verbs
9.5 The modal construction
At a systemic level the acquisition of the modal construction can be described as a shift from a stage where “root infinitives” in various (semantically) modal contexts are a dominant utterance type (cf. Josefsson 2002) to a stage where the infinitive appears as verbal complement to actual modals. It is meaningless to try to establish when the target modal construction (modal + verbal complement in the infinitive) replaces an earlier system with root infinitives. The reason for this is three-fold: 1) the infinitive in Swedish (for instance kasta ‘throw’ and springa ‘run’) has the same structure as the most frequent verb form for both Hugo and Bruno up to approximately 1;11, 2) this form is not only an infinitive (it can also function as an imperative and as a present-tense form lacking the final suffix –r), and 3) root infinitives definitely have other functions than appearing in modal contexts (cf. Christensen 2003a). There are simply too many verbs that resemble infinitives and
appear in various functions. In comparison, the emergence of the perfect construction can be described as a shift from “root supine” to full perfect, since the supine is unique to the perfect construction. (This, too, will be addressed in next section.)
The emergence of the modal construction is definitely covered best in the files on Hugo. During two months (2;0 and 2;1) he produced several instances of the non- target-like form “villa”, a combination of the modal vill ‘want to’ and the a- morpheme, which is typical of the Swedish infinitive. A total of seven entries in the diary, all from 2;0 and 2;1, include a report about the “villa” form. For instance, it is registered five times in combination with inte ‘not’: “villa inte sova egen säng” [want.to not sleep own bed] ‘I do not want to sleep in my own bed’ (H 2;0) and “villa inte dokton” [want.to not doctor.the] ‘I do not want to go to the doctor’ (H 2;1). The combination “villa aa” [want.to have] is reported twice. At 2;0 it is noted in the diary that the string “villa aa x” [want.to have x] (where x stands for a noun) is, at the time, heard several times a day as a replacement for the earlier string “vill x” when Hugo wishes to have a certain object. (There is however no earlier entry reporting or quoting the formula “vill x”.) At 2;1 he uses “villa aa” several times when offering me a ball to play with. In the diary it is commented that it “almost sounds like ‘vill du ha?’ [want.to you have]”, which is a target-like way of expressing ‘do you want to have it?’. When the form “villa” is followed by “aa” [have] it is for obvious reasons hard to tell if the verb actually is expanded with an –a or not, but when it is followed by the adverbial inte ‘not’ it is more tempting to interpret the final –a as the infinitival morpheme. My hypothesis, therefore, is that villa is an over-regularisation into the most typical Swedish morphological infinitive form (which is disyllabic, ending with an –a) during a period when almost all of Hugo’s verbs appear in this form. Note that modals in the infinitive do occur in adult Swedish, but not frequently. The target-like infinitive vilja ‘want to’ is not registered during the investigated period in either boy. Actually it is neither found in later recordings, which have been checked up to the age of 3;11. The form villa also seems to represent a certain stage in the syntactic development of the modals, cf. below.
Syntactically the modals show a quite distinct chronological development. Not surprisingly, the pattern stands out more distinctly for Hugo than for Bruno, but the data on Bruno does at least not contradict that for Hugo.
reported in the diary at 2;0 that vill ‘want to’ has been registered “for a long time” and that also kan ‘can’ and måste ‘must’ are now heard regularly. Actually it is not stated explicitly whether the verbs have been registered in one-word-utterances or in more elaborated linguistic strings, but I suspect that it has been in one-word-utterances since no entries quote longer segments. For Hugo “kan inte” ‘cannot’ and “vill inte” [will not] ‘do not want to’ are said to be common utterance types at 1;9, with the additional information that “kan inte” has been observed “for a long time”. A similar utterance type is “kan fälv” [can self] ‘I can do it by myself’ (H 2;1). As mentioned earlier, vill + a noun is also registered (but reported at a later point in time) during this early period.
At 2;0–2;1 the form villa is registered seven times in the diary for Hugo, cf. above. This form is not only morphologically interesting, it also seems to represent a mediating level syntactically, since it is the first modal to be constructed with a subordinate main verb but without a subject. Examples: “villa inte sova egen säng” [want not sleep own bed] (H 2;0); “villa aa [x]” [want have x] (H 2;1). The form is also found without a verbal complement, but instead with a nominal complement, as in “villa inte Gurun, Lena” [want not Gudrun, Lena] (probably meaning ‘I do not want to go to Gudrun, I would like to go to Lena instead’; Gudrun was Lena’s substitute) (H 2;0); “villa inte dokton” [want not doctor.the] (H 2;1).
For Bruno the only utterances that would be possible to interpret as representatives of a mediating level of development are “ja kunde” [I could] (B 2;2) and “ja kunde, mamma” [I could, mummy] (B 2;2). This utterance type is reported as being heard regularly when Bruno succeeded in doing what he intended – without help from anyone. Here we see a subject but no main verb. Both utterances are however target- like, since the main verb can be left out in adult Swedish in this specific context.
The third step in the development is the full modal construction, i.e., a complete modal sentence consisting of both a subject and a complement verb. For Hugo the transition to this stage seems possible to pin down to a very limited period lasting from 18 to 26 March 1994 (when he was in the middle of 2;1). On 18 March he uttered his first complete modal sentence, “kom ka vi tända” [come shall we turn.the.light.on] ‘let us turn the light on’, which seems to be holophrastic – this is probably exactly what I usually said to him in the morning. On 19 March he uttered the sequence “villa aa?” repeatedly, which I, as mentioned previously, at the time believed to be very close to the question vill du ha; this is otherwise the last instance
of the “villa”-form. The same day he uttered the complete modal sentence “gubben måste sitta där” [man.the must sit there], which I do not regard as holophrastic. The next day “ja ska baan dicka, sa ja, ja ska baan dicka” [I will just drink, said I, I will just drink] is registered. This utterance seems to have been influenced by his peers at his child minder’s since the sequence “ja ska bara” is typical for child Swedish and the segment “sa ja” is otherwise found in utterances that seem to have been picked up at his child minder’s. On 21 March the odd “vill inte sovit” [want not slept] is registered, with a supine as the main verb and no subject. (Note that sova is a verb that often appears in the supine for Hugo.) This is the last true instance of a modal sentence without subject.
On 23 March the first instance of an utterance with the general pronoun man ‘you’ and kan ‘can’ is registered: “man kan tycka knappen, man kan skua dä, man kan öppna, tänga” [you can press button.the, you can turn there, you can open, close] (when examining the washing machine). The same day another pronoun subject is found: “du måste nalle” [you must teddy], the lack of a main verb here is target-like. The next day we find yet another subject: “vi ska leka” [we shall play], and on 26 March we find an inverted word order for the first time: “nu ska *Hugo sjunga, nu ska *Hugo telle” [now will Hugo sing, now will Hugo narrate]. After this period there are no instances of modals registered without subjects – apart from one entry noting that the modal ska is frequently heard. The only instances registered without subordinate main verbs are target-like (and rare), for example “Mattin inte skulle de” [M not would that] (the word order is non-target-like). One may conclude that Hugo, over a period of eight days, established the complete modal sentence with various subjects and various main verbs in his productive language, and also uttered examples of target-like subject-verb inversion (i.e., XVS word order).
Bruno’s first complete modal sentence is “ja kan inte sätta på locket, mamma” [I can not put on lid.the] ‘I cannot put on the lid’ (B 2;2). My estimate is that the utterance is not holophrastic; there is no given master for it in adult Swedish, even if it is fully target-like. Although we cannot follow the emergence of the full modal construction for Bruno, it seems to be present at 2;2. From 2;3 there are for Bruno a total of seven utterances with the full modal construction.
To sum up, the modal construction has the following developmental stages: A) “bare modals”, i.e., modals with no subordinate verb and with no subject (but
complement in the infinitive, but still with no subject; C) Full modal construction consisting of subject, modal and verbal complement. Modals with heavier semantic content (vill ‘want to’, kan ‘can’, måste ‘have to’) show up earlier than the intentional future construction (with ska ‘shall/will’).
As has already been mentioned, modals without verbal complements are sometimes fully target-like. During the very last month discussed here (i.e., 2;5 for Hugo and 2;7 for Bruno) the typically Nordic construction with a modal + a directional adverbial (where a directional verb can be said to be left out) is found: “ja vill inte till Serstin” [I want not to Kerstin] ‘I do not want to go to K’ (H 2;5); “ska vi på dagis” [shall we on day.care.centre] ‘are we going to the day care centre’ (B 2;7). Both utterances are target-like. One can conclude that this more peripheral use of the modals emerge later than the main use of the modals.