146 Małgorzata Krzek
of generic exclusive.^18 In this respect, the Polish null subject DP is similar to English
one, which ‘generally ‘but not always’ includes the speaker as well as the addressee’
(Moltmann 2006: 3).^19 This appears to suggest that the binding of the feature [indefi-
nite] by the generic operator is not enough to ensure a generic inclusive interpretation.
The sentence in (26e) provides further support for this hypothesis. Firstly, for all of my
informants, the null subject can only be interpreted as exclusive. In other words, they
do not perceive themselves as being among a possible group of referents of which the
main verb is predicated.
This seems to suggest that, at least in this case, the inclusion of the speaker is a
matter of choice and that grammatical means present in the structure are sometimes
not enough to enforce inclusive reading. As such, this observation appears to provide
support for the hypothesis that pronouns are constructed in Narrow Syntax. Another
point to note about this sentence is that the null pronoun is ambiguous and can be
interpreted as either arbitrary or generic exclusive. This suggests that in this particular
case it is the domain of the pronoun that is constrained by the adverbial, and different
interpretations depend on how the adverbial w Paryżu ‘in Paris’ is interpreted; that
is, whether one considers this adverbial as restricting the possible set of referents to a
small group or not.^20 The last point that needs to be mentioned here is the agreement
with predicative complements. It is assumed here that the mechanics is the same as in
the case of interpretation. Namely, the properties (phi-features) of speech participants
get encoded in the CP domain from where they bind the features of pronouns. Those,
in turn, provide values for predicative complements. Therefore, the marking on the
predicative complement in the impersonal SIĘ construction will change depending on
the referents’ feature specification (as illustrated by the sentences in (26) above). On
the other hand, the null pronoun in the –NO/–TO construction only triggers mascu-
line, plural marking on the predicative complement regardless of the feature specifi-
cation of possible referents. According to Marek Świdziński (p.c.), sentences such as
the following (in (27)) are infelicitous when a referent is a group consisting of females- See D’Alessandro (2004, 2008 ) for a similar observation. D’Alessandro (2004, 2008 )
assumes that generic interpretation is not specified for inclusivity. Therefore, inclusivity must
be enforced by other means. - See Dalmi (2013) for a discussion of the interpretation of impersonal constructions in
Hungarian and interactions between generic inclusive DPs and temporal/locative adverbials. - The figure in (i) below illustrates how the binding of the features on pronouns might
look like. It is worth pointing out that logophoric features themselves should be considered as
complex ones, that is consisting of such features as number and gender. These subfeatures of
the logophoric features bind some of the [Class] and [Group] subfeatures. Features that are
not bound will be deleted.