144 Małgorzata Krzek
Because of that, I will opt for adding features [generic] and [arbitrary] to the geom-
etry. The addition of the feature [non-specific] appears to be a necessity as there is a
perceptible difference between [plural] on the one hand and the subfeatures of the
feature [non-specific] on the other. Namely, the features [generic]/[arbitrary] can
describe a group that is definite or indefinite but it can never be specific if specificity
is understood as identifiability and familiarity. This is, I guess, not the case for the
feature [plural], which I assume can pick out referents that can be either indefinite
or definite but they will always be specific in the sense that they are familiar and
identifiable.^17
Generic/arbitrary interpretations are dependent on the interplay of various con-
straints imposed by possible restrictors and binders. These elements are (a) logophoric
features (ΛA, ΛP) that bind the [speaker] and [addressee] sub-features on a pronoun
and (b) temporal adverbs such as zwykle ‘usually’, zawsze ‘always’, często ‘often’, nigdy
‘never’, etc. These adverbs, I assume, bind the [indefinite] feature of a pronoun. It is
possible because these temporal adverbs are quantificational (Q-adverbs) and in this
respect very much like generic operators (Dobrovie-Sorin 2001). If Q-adverbs are not
overtly present, then either the covert generic or existential operator will bind the sub-
feature [non-specific]. For a detailed discussion of the problem see Krzek (2013). For
the time being, it is sifficient to observe that the introduction of the feature [generic]
and [arbitrary] to the feature geometry will allow us to capture the difference between
generic exclusive and arbitrary readings.
As for locative frames, such as w Polsce ‘in Poland’, w ogrodzie ‘in the garden’,
I assume here that they are frame-setting modifiers (Maienborn 2001) that restrict the
space and time of the eventuality (also known as domain adverbials, in Bellert’s (1977)
terminology). It is to be noted, however, that some of these adverbials will be more- Pronouns can be definite, that is having their phi-features specified, and still refer to
some non-specific group of people. Consider the second sentence in (i) below:
(i) W Średniowieczu rzadko jedzono produkty zawierające
in Middle-Ages rarely ate.ipfv products containing
węglowodany. Dużo częściej jedzono mięso.
carbohydrates more often ate.ipfv meat
‘In the Middle Ages [people] rarely used to eat products containing
carbohydrates. More often [they/people in the Middle Ages] used to eat meat.’
The referent of the agent in the second sentence is easily identifiable as people who lived
in the Middle Ages. This appears to suggest that definite but non-specific elements can be
topics. The same is true about DPs (I would like to thank Peter Ackema for pointing this out
to me). These can also be definite and receive a generic interpretation. This is illustrated by
the following example:
(ii) The horse is a noble animal. (Peter Ackema, p.c.)