The syntax of inflectionverb cannot escape the VP because of its -assigning properties, pointing to the fact
that aspectual auxiliaries and copular be, which do not assign -roles, can appear in I
(Pollock 1989).
But from our perspective, these elements appear in I by being inserted there and do
not undergo movement at all and so there may be other reasons for the fact that they
behave differently to main verbs. It is also not entirely clear why the verb can move
within the vP, sometimes through as many as three light verb positions and not have
any trouble with its -assigning properties. Something rather stipulative and
ultimately circular has to be claimed to try to account for this fact. For example, we
might assume that the inflection has some property, which light verbs lack, that means
that if a thematic verb moves to I it cannot assign its -roles. Often it is claimed that
the inflection is ‘too weak’ to support the verb’s -assigning requirements. But the
weakness of an element only correlates with the ability of the verb to move to that
element, which is the very reason for proposing the notion in the first place!
Before trying to solve these puzzles, one more mystery should be introduced. Our
assumptions have been that auxiliaries are inserted into a structure to support bound
morphemes when the verb is unable for one reason or another to do so. Obviously a
free morpheme does not need supporting either by the verb or an auxiliary. This would
predict that when the inflection appears as a free morpheme, i.e. a tense or the
infinitival marker, there will be no need for an inserted auxiliary to accompany an
aspectual morpheme. But this prediction seems to be false:
(38) a he will have gone
b she might be worrying
c for you to be seen here would be disastrous
The obvious question is what are these auxiliaries supporting? Note that any element
that appears after a free inflectional element is always in its base form. Thus, either the
auxiliaries are supporting nothing, which throws doubt on their treatment as inserted
empty elements, or they are supporting a null morpheme. The latter assumption allows
us to maintain our approach but it raises the subsequent question of what this
morpheme is.
The facts concerning this morpheme are that it is only present when there is a free
inflectional element and the morpheme always follows the inflection.
(39) a he will leave-∆
b they must have-∆ left
c we might be-∆ leaving
d to be-∆ seen
When there is a bound inflectional element, i.e. a tense morpheme, the null morpheme
is not present:
(40) a we did have-∆ gone we had gone
b they did be-∆ going they were going
c *I did leave-∆ I left