Abstract for Francesca del Gobbo's talk

On Chinese Appositive Relative Clauses
Francesca del Gobbo (UQAM)
May 22, 2007 (Tuesday)
6:30 PM - 8:00 PM; Room 7102, The CUNY Graduate Center

Download the abstract in PDF

Cinque (2006), on the basis of evidence from Italian, shows that there exist two types of Appositive Relative Clauses (henceforth, ARCs): integrated and non-integrated ones. English can only have non-integrated ones. In this talk I claim that Chinese ARCs are adjectival (Del Gobbo 2004, 2005) in the sense that they can only be integrated ones, and I propose that the typological difference is due to the type of relative pronoun and to linear order.
Cinque (2006) proposes a unified syntactic analysis of RCs, whereby RCs originate prenominally; and both matching and raising are needed:
(1) [DP the [CP that [[IP John bought [dP2 expensive book]] [dP1 [[AP expensive] [NP book]]]]]]
From the basic structure represented in (1), a raising structure is obtained by raising dP2, and by PF deletion of dP1; a matching structure, by raising of both dP1 and dP2 and PF deletion of dP2.
Following Cinque (2006), I propose the following matching structure for Chinese RCs:
(2) [DP D [CP3 [CP1[IP wo xihuan tdP2] de tIP [dP1 xiaohuozi]] [CP2 [dP2 xiaohuozi] tCP1]]]
I like DE guy guy
‘the guy I like’
The structure in (2) is obtained by first raising dP2 (to account for island effects), then raising CP1. The absense of reconstruction effects is captured by the fact that the ‘head’ is not directly linked to the trace within IP. I assume that PF deletion of dP2 is allowed by cyclic c-command (Huang 1982). For cases of raising, I instead propose the following structure:
(3) [DP D [CP3 [CP1[IP ta chi tdP2] de tIP [dP1 cu]] [CP2 [dP2 cu] tCP1]]]
he eat DE vinegar vinegar
‘his jealousy’
The structure in (3) is obtained by raising dP2 (to account for island effects). After dP2 raises, it c-commands dP1, thereby allowing its deletion at PF. Only afterwards, CP1 raises. The presence of reconstruction effects is captured by the fact that the ‘head’ is directly linked to the trace within IP.
Interestingly, the main differences between English (representing post-nominal RCs) and Chinese (representing pre-nominal RCs) are:
1. dP1 never raises in Chinese, the remnant CP1 does;
2. in Chinese, dP2 is never a pronoun.
Notice that crosslinguistically, no prenominal RC allows pronouns. If the structure and derivation proposed here for prenominal RCs are correct, we can explain the unavailability of pronouns in Chinese RCs (and typologically in all prenominal RCs) by assuming that relative pronouns need to be c-commanded and bound by the ‘head’ of the relative clause (see Fukui and Takano 2000, Cinque 1982, among others). In fact, in the two structures postulated here for Chinese (i.e., those in (2) and (3)), this is not possible. This will also have some consequences for ARCs.
I claim, in fact, that non-integrated ARCs are the canonical or default ones: they are parentheticals, so, at a certain level, they are independent sentences. All the properties we are aware of follow from this. I also claim that a canonical ARC requires the following three conditions to be met:
1. the link must be spelled out, i.e. the pronoun must be overt, it can’t be an operator (so, in our terms, it can’t be a raised nominal);
2. the pronoun must be of the type also used in interrogatives (see Cinque 1982);
3. the pronoun must follow its antecedent, as a linear order requirement imposed by discourse.
If one or more of the above conditions isn’t met, the language resorts to other strategies.
Extending the proposal put forth for restrictives to ARCs in Chinese, we obtain the following structure:
(4) [DP D [CP3 [CP1[IP wo xihuan tDP2] de tIP [DP1 Zhangsan]] [CP2 [DP2 Zhangsan] tCP1]]]
I like DE Zhangsan Zhangsan
Lit. ‘Zhangsan who I like’
The main difference with restrictive matching RCs is given by the fact that here it’s the entire DP that raises, not the dP.
Notice that given (4), Chinese doesn’t satisfy any of the conditions above, as the relative pronoun is not spelled out, it is not of the type used in interrogatives (which btw stays in situ), and it doesn’t follow its antecedent. It follows that for its appositives Chinese is forced to use the integrated strategy.