Jaume Solà i Pujols - Departament de Filologia Catalana ...

The suggestion in d) that non-NSLs have to choose 0.a) (at least as a preferred option) is,
I think, of historical interest: Languages loosing the Null Subject status quickly develop
postverbal negation (this is what happened in the passage from Old French to Modern French,
and from Old English to Middle English:
(35) Idealized data:
a. Old French: Je ne mange
I not eat
b. Modern French: Je (ne) mange pas
c. Old English: I no come
I (not) eat not
d. Middle English: I (no) come nought
These changes are contemporary of the loss of the Null Subject status. Since they are
rather spectacular changes ( they are not trivial reanalysis processes), it is plausible that they
were forced by UG. Our suggestion is that a non-NSL has to take the unmarked option of
lowering ΣP and promoting ModalP as the top FC in order to make the top FC an appropriate
host for subject-AGR, as explained in d) above. The reason why subject-AGR has to be the host
of the top FC is probably that this is the only way for left dislocated or pro-resumed DPs to be
easily reanalysed as AGR-identifiers when the language loses the Null Subject ability.
1