6.2.2 Do-insertion
The use of have and be as supporting auxiliaries is therefore associated with the appearance of the aspectual morphemes whose presence necessitates the use of the auxiliary by ‘tying-up’ the verb so that it cannot support any other morpheme. The use of the dummy auxiliary do however, is a little different as it is not associated with the appearance of any aspectual morpheme and indeed cannot be used in the presence of one:
(21) | a | he did not arrive |
b | he had not arrived | |
c | *he did not have arrived |
What determines the use of the auxiliary here? Obviously the verb is unable to support the inflection in this case, but this does not seem to be because it already supports another morpheme. In fact the verb is in its base form and there is no reason to think that there is any other verbal morpheme present. (21a) is simply the negative version of he arrived. Apparently it is the negative that blocks the verb from moving to support the inflection. To gain some understanding of what is going on here we need to briefly examine another kind of head movement which we will more thoroughly discuss in the next chapter. In the formation of certain questions an auxiliary verb is moved to the other side of the subject:
(22) | a | Denise will dance | will Denise dance? |
b | Tim is tall | is Tim tall? |
As we can see, both modal and aspectual auxiliaries can undergo this movement process. The observation of interest to us is what happens when there are more than one auxiliary:
(23) | a | Graham could be gardening |
b | could Graham be gardening? | |
c | *be Graham could gardening? |
Apparently, when there are more than one auxiliary, the first one is chosen to move. The reason for this seems to be that moving the first auxiliary involves a shorter movement than moving the second:
(24) | a | |
b |
Travis (1984) proposed that this phenomenon can be explained by a restriction on head movement which prevents one head from moving over the top of another:
(25) | the Head Movement Constraint (HMC) | |
a head must move to the next head position |
The reason why (23c) is ungrammatical, then, is that if the aspectual auxiliary moves in front of the subject, it has to move over the modal. Whereas if the modal moves, it crosses over no other head. Now consider the case of verb movement in the presence of not:
(26) | a | = he rang the bell | |
b | = *he rang not the bell |
The movement represented in (26a) appears to be grammatical whereas that in (26b) is ungrammatical. Again the difference between the two is that the grammatical movement is shorter. But if we want to use the HMC to account for the phenomena, it must be the case that the negative is a head as it is moving over this element that causes the problem. But, what kind of a head is the negative? It is situated between the inflectional element and the v/VP:
(27) |
We know that the inflectional element takes a v/VP complement and therefore that the negative must be either V or v. As the complement of the negation is a v/VP it follows that the negative must be v, a light verb, as main verbs do not have verbal complements. Thus the analysis is:
(28) |
Accepting this, we can account for the insertion of dummy do. The verb will not be able to move to inflection without violating the HMC. Apparently in English, the negative is not the sort of verbal element that can support tense and hence the only option available is to insert an auxiliary. As there is no aspectual morpheme to deem otherwise, the inserted auxiliary will be do:
(29) | ||
‘the glass did not shatter’ |
Note that the inability of the negative to support the inflections is a language specific property and there are languages where this is exactly what happens. For example, Finnish negation shows the same agreement morphemes as its verbs do and in the presence of negation the verb does not inflect for agreement:
(30) | menen | – | en mene | ||
go1.s. | not1.s. go | ‘I go/I don’t go’ | |||
menet | – | et mene | |||
go2.s | not2.s. go | ‘you go/you don’t go’ | |||
menee | – | ei mene | |||
go3.s | not3.s. go | ‘he/she goes/he/she doesn’t go’ | |||
menemme | – | emme mene | |||
go1.pl. | not1.pl. go | ‘we go/we don’t go’ | |||
menette | – | ette mene | |||
go2.pl | not2.pl. go | ‘you (lot) go/you (lot) don’t go’ | |||
menevät | – | eivät mene | |||
go3.pl | not3.pl go | ‘they go/they don’t go’ |
Because of its behaviour, the Finnish negative element is often called the negative auxiliary or even a negative verb. Moreover, in other languages the negative element surfaces as a bound morpheme on the verb, a situation very similar to the analysis we have given the aspectual markers in English. This is exemplified by the following Choctaw and Japanese sentences:
(31) | ak-Ø-pi-so-tok | |
1s-3s-see-not-past | ||
‘I didn’t see it’ | ||
(32) | watashi-wa yom-anakat-ta | |
I read-not-past | ||
‘I didn’t read’ |
Besides the bound morpheme status of the negative, these languages differ from English in that verbal stems are allowed to support more than one bound morpheme and hence there is agglutination: complex words being formed from a series of inflectional morphemes. The point is that in these languages the negative element behaves like we have seen certain English light verbs do and hence they offer support for the suggestion that the negative can be analysed as a light verb.
Note that the presence of the negative will not affect the use of aspectual auxiliaries as these are inserted into the inflection position rather than moving to it:
(33) | ||
‘the glass has not shattered’ |