152 Małgorzata Krzek
would only be available for existential operators but not for universal ones, as is argued
by Diesing (1992).^27
Crucially, SIĘ is devoid of the referential index [iR], and as such it cannot be
bound by either an A-topic or any of the speech features (ΛA, ΛP). It does not block,
however, the speech features from binding the [participant] subfeatures on pro. There-
fore, an inclusive reading is still possible. Because SIĘ does not have [iR], the phrase it
heads will never be able to have a specific and definite interpretation, namely that of
the 1st/2nd/3rd person pronoun. It will only be able to get an indefinite reading. This
will allow us to explain the difference between the sentences in (31).(31) a. Je się zupę.
eats się soup.acc
‘[One] eats soup.’
b. Je dużo zupy.
eats a-lot-of soup.gen
‘[He/She] eats a lot of soup.’In (31a) pro is merged as a complement of the SIĘ functional head. As SIĘ does not
have [iR], the speech features, (ΛA, ΛP) in CP will not be able to bind it. They will
only bind the [participant] subfeatures of pro and the result will be the inclusive
reading. In (31b), on the other hand, pro will be merged as a complement of the null
D head that carries [iR], and as a result it can be bound by an A-topic in CP and
receive a definite interpretation. We can then say that SIĘ is like the D head of the DP
projection, or more specifically, that they are elements of the same type. However,
unlike D with [iR], which is associated with definiteness, SIĘ is associated with the
absence of it.
Additional evidence that appears to support the lack of referential index [iR] on
SIĘ comes from the observation of binding properties of null impersonal pronouns.
Consider the sentences in (32) and (33).- As rightly observed by an anonymous reviewer, it appears that the movement out of
the VoiceP invokes look-ahead since the operator in the CP has not been merged yet. In the
framework assumed here, this type of movement cannot be motivated by the case assignment
considerations either, as it is argued that T does not assign case. One of the possible motiva-
tions for the movement of pro out of the VoiceP would be that all DPs have to leave their
first-merged positions. This is, however, just a stipulation, and a more in-depth investigation,
which is outside the scope of this paper, is necessary to discover an independent motivation
for this movement.