Why the Blocking Effect?

Haihua Pan

CTHPAN@CityU.Edu.HK

Several proposals have been made about ziji, whose antecedents, though subject to the so-called blocking effect (Tang 1989), can be outside its Governing Category (GC). The LF-movement account claims that ziji is covertly bound to a local antecedent, observing Chomsky's (1981) Binding Condition A (BCA). The parameterization approach suggests to parameterize GC, or to relativize SUBJECT, so that ziji can be bound in a domain larger than a clause. The recoindexation account allows ziji to be recoindexed, in order to satisfy BCA. The other-factor approach argues that ziji is not constrained by BCA, but by discourse factors such as thematic prominence, logophoricity, emphasis, or perspectivity. Although there are pros and cons for these proposals, the first three fail to explain the blocking effect, especially (1) and (2) below, while the fourth cannot explain the facts that long-distance (LD) bound ziji observe the blocking effect and exhibit subject orientation.

(1)  Ni i shuo-guo tade j shu hai-le ziji i/j ma?
     you say-Guo     his book hurt-Perf self Q
'Have you ever said that his book hurt him/you?'
(2) John i shuo nide j pengyou hai-le ziji *i/j .
            say your friend hurt-Perf self
'John says that your friend hurt him.'
Baoyüi renwei [woj de xuesheng]k bu xihuan ziji*i/*j/k.
          think I DE student not like self
'Baoyü thinks that my students does not like himself.'



Woi bu xihuan Billj guan zijii/j de shi.
  I not like    interfere self DE matter
`I don't like Bill interfering my (own) business.'



Woi zhidao Billj de shu hai -le zijii/j.
 I know         DE book hurt-Perf self
`I knew that Bill's book hurt me/him.'



Johni zhidao Billj zai wo/nik jia xi ziji*i/j/k de zhaopian.
      know         at I/you home develop self DE photo
`John knew that Bill was developing his/my/you picture(s) at my/your home.'

A careful examination of the usage of ziji and its compound forms, based on textual search of large corpora, reveals that (a) only LD bound ziji exhibit subject orientation, and (b) unlike the symmetrical unlike-person subject blocking claimed in the literature, only intervening 1st/2nd person NPs (both subjects and non-subjects) can block ziji from being LD bound. The examination suggests that self-ascription and discourse prominence (DP) play an essential role in Mandarin reflexive interpretation. Thus, contrastive and non-contrastive reflexives should be differentiated: while the former are constrained by DP, the latter should be further divided into locally and non-locally bound reflexives. The thesis of this paper is that non-locally bound ziji is a de se anaphor, borrowing Lewis' (1979) terminology, and thus can only be bound by the carrier of beliefs, which requires at least beliefs de dicto. This thesis provides a natural explanation to the behavior of ziji. LD bound ziji is possible, since 3rd person NPs optionally introduce beliefs de dicto and thus can satisfy the condition of ziji. The blocking effect is explained away by the fact that 1st/2nd person pronouns obligatorily introduce beliefs de re, non-de dicto beliefs, and thus always violate the condition of ziji. Subject orientation is observed because subject is usually the only carrier of beliefs.

I will also argue that non-locally bound ziji is not an instance of intensive pronouns, as suggested in Baker (1995) for (British) English himself, and claim that the partition of morphological reflexives is not a binary one between locally bound and non-locally bound reflexives, but a tertiary one, as suggested above.