2. SUBJECTS
2.3. Preverbal Subjects
2.3.4. Left-dislocated Subjects and the Null-subject Parameter
definiteness, Person and Number. In Italian, an example that corresponds to the one in (99a) is ungrammatical:
(104) *le donne denunziamo le ingiustize [the women denounce-1stpp the injustice]
b. ?*o anthropos irthe ithele na sou milisee o anthropos bou ithele na sou milisee irthe c. a man came that wanted to talk to you a man that wanted to talk to you came
Nevertheless, as in (106), extraposition in English is blocked when the head of the relative clause is a definite NP):
(106) *the man came that wanted to talk to you182
These authors then show that there is empirical evidence which supports an analysis of preverbal subjects in null-subject languages as CLLD constructions.
They also propose that agreement morphology in null-subject languages includes a nominal element according to classical tradition in the PP literature. They offer a standard minimalist account of the properties of null-subject languages as follows.
In these languages, V-to-AgrS raising is forced by the presence of a [+strong]
nominal feature in this inflectional head that forces overt raising of a syntactic element with a nominal feature in order for the derivation to converge, as in (107):
182 Example taken from Olarrea (1996, 128).
V IP
I’
I VP V’
[+strong] [+strong]
(107a)
V
*IP I’
I VP V’
(107a)
Since V presents this feature, head movement in (107a) and not phrasal movement in (107b) can satisfy the EPP, i.e. can break the [+strong] nominal feature of the highest inflectional head. Obligatory V-to-I raising in null-subject languages is, therefore, explained and the pronominal properties of I in these languages is accounted for. Moreover, it is argued that head movement should be considered less costly for the computational system since it does not expand the phrase marker, that is, since it does not create new structure, and is, in this sense, similar to covert movement.
2.3.4.1. Preverbal Subjects and Interrogative Sentences: Pro and Lexical NPs If we assume that Spanish preverbal subjects are CLLD constructions, the question that arises is why preverbal subjects are not allowed in Spanish interrogatives unless the wh-phrase is generated in [Spec CP]. Olarrea (1996) speculates that IP-adjuncts (i.e. preverbal subjects) block covert I-to-C movement.
As a result, preverbal subjects are not allowed in Spanish when there is a fronted wh-phrase in [Spec CP]:
CP wh-phrase
Spec C’
C IP I’
VP
...
I
preverbal subject
*qué Marta quiere
(108)
Spec
A representation such as the one in (108) would be ungrammatical, therefore, since the preverbal subject (Marta) is blocking I-to-C movement. I cannot raise covertly to C and the required configuration for the satisfaction of the wh-criterion is not met. The derivation will crash at LF due to the presence of a preverbal adjunct, independently of the argumental/non-argumental status of the wh-phrase.183
Interrogative sentences oppose relative ones in this respect. The presence of left-dislocated constituents (preverbal subjects, for instance) is grammatical in the case of relative clauses, since they do not present I-to-C movement.
This proposal also predicts that the wh-elements which are base-generated need not be licensed by covert I-to-C and therefore will allow the presence of preverbal subjects in affirmative sentences. At the same time, the analysis of preverbal subjects as base-generated adjuncts to the maximal inflectional projection coindexed with an empty pronominal in argument position correctly predicts the different distribution of preverbal subjects and empty pronominal subjects in Spanish in interrogative sentences. Consider the examples in (109):
(109) a. ¿qué pro quieres?
[what pro want-2ndps]
[what do you want?]
b. *¿qué Juan quiere?
[what Juan want-3rdps]
[what does Juan want?]
While an empty subject in an interrogative sentence is allowed, as (109a) shows, the presence of a preverbal subject (Juan) in (109b) renders the sentence ungrammatical, as has been previously said.
If we analyze both the preverbal subject and pro as occupying the same position, i.e., the canonical position [Spec AgrSP], an accurate account for the contrast present in the previous examples (109a) and (109b) would still need to be given. The contrast is directly derived from the fact that empty subjects are in [Spec VP] prior to Spell-Out and by the fact that preverbal subjects and empty pronominals never occupy the same position in Spanish. The tree diagram in (110) captures the position of preverbal subjects and empty subjects that we have been discussing:184
183 Suñer’s (1994) and Olarrea’s (1996) analyses contrast in slightly different ways with that of Torrego (1984). For Torrego, VSO order in Spanish is derived from a basic SVO order by a rule of V-fronting, so that the verb moves to a position higher than IP.
184 As Olarrea (1996) points out, this is true only of what is referred to as standard Spanish, as opposed to Caribbean Spanish. In the latter, the presence of a pronominal element that precedes the verb in an interrogative is perfectly grammatical, as the example in (i) shows:
(i) ¿qué tú quieres?
AgrSP ... AgrSP
AgrS’
VP
V’ ...
NP
Spec
AgrS
Spec
preverbal subjects
empty subjectsi
quiere viajar ti
pro Marta
[Marta pro wants to travel]
(110)
Preverbal subjects and empty subjects never occupy the same position in Spanish and this is how the contrast in the previous examples (109a) and (109b) is explained: empty subjects rather than preverbal ones are accepted in interrogative clauses.