I'm reading the paper Constraints on Null Pronouns by Speas, in which the author defines two constraints for a cross-linguistic OT analysis of the occurrence of null pronouns across languages:
Control(CTL): A featureless pronoun must be coindexed with the closest c-commanding nominal element in its C-Domain.
Binding Theory Principle B (BTB): A pronoun must be free in its Binding Domain.
and the author also defines what she means by "C-Domain":
C-Domain: α is the C-Domain for β iff
- α dominates β
- α dominates some set of checked feature(s) γ and
- γ c-commands β and
- no other set of checked features c-commands β
I assume by the "Binding Domain" of an α the author meant the Minimal Governing Category for α, which consists of:
- maximal projection that contains α,
- a SUBJECT accessible to α, and
- the governor of α.
My understanding is that the two constraints are close opposites to each other (one asks featureless pronouns to be bound, one asks pronouns to be unbound), and the important point of the author's argument is that different languages allow different occurrences of the null pronoun since they rank these two constraints (and 4 others she mention) differently.
However the issue is whether this oppositeness always holds. I think that depends on whether the C-Domain of a null pronoun would always match with its Binding Domain, but I cannot answer my own question. Any thoughts?
By the way, the author considers the null pronoun to be a featureless pronoun. Thus null pronouns are subject to both of the two above constraints.