• No se han encontrado resultados

Programa:44 GESTION ELECTORAL

So far, the data seem to indicate that in Russian (and in Slavic in general) the phase head undergoes v-to-Asp movement. Assuming Phase Extension hypothesis (den Dikken 2007; Gallego 2007, 2010, among others), such movement results in creation of a hybrid head v/Asp and extension of the vP phase to v/AspP, as

originally suggested by Dyakonova (2009) for Russian.13

Here I want to extend the logic of Gallego’s (2007, 2010) analysis to Slavic by proposing that in these languages v bears an unvalued instance of Aspect feature ([uASP]), which is checked through v-to-Asp movement. Moreover, assuming Svenonius’ (2004b) idea that a phase can be spelled out only when all uninterpretable features on its head are checked and deleted, we predict that the

vP phase cannot be transferred until [uASP] is removed from v. In other words,

Transfer of v is delayed until v-to-Asp head movement takes place, assigning the phasal status to the AspP projection.

13 See Bošković (2014), who also argues for the phasal status of AspP in the extended VP

domain. Although such status is reached for reasons independent from Phase Extension process, this analysis supports our intuition.

121

Applying the phase extension mechanism to Slavic, I argue that head movement of v gives rise to a parameterized phase extension of vP into v/AspP, as represented below: (59) AspP Asp vP YP v’ v VP [uASP] VP XP v/AspP v/Asp vP [uASP] YP v’ tv VP VP XP COMPLEMENT DOMAIN EDGE DOMAIN

As discussed in den Dikken (2007) and Gallego (2007, 2010), a natural effect of the phase extension is the consequent extension of the phase complement domain, a point which is going to be crucial for our purposes. Recall that in accordance with PIC, only the head H and the edge of a phase HP are visible for further operations, while the complement of H becomes opaque, by being transferred. However, if the phase HP extends to ZP, what used to be the edge of HP turns into the domain of the newly constructed phase ZP. As a result, all syntactic objects with unvalued feature(s) are forced to escape the domain of ZP, otherwise the derivation will crash. In other words, such elements move to the edge of ZP.

Applying this logic to Slavic, I suggest that after phase extension the edge of vP turns into the domain of AspP, as illustrated in (59). Subsequently, all potential goals or elements with a formal imperfection must be removed from the edge of vP to the edge of the newly constructed phase, AspP, in order to be visible for further syntactic operations. Within our proposal, this is exactly what forces movement of all wh-items in a multiple question to the preverbal area.

Consider first the predictions of our analysis for Bulgarian multiple

wh-questions, as in (60):

(60) a. [QP Koj1] [QP kogo2] e celunal? (Bulgarian)

who.NOM who.ACC kissed

‘Who kissed whom?’

b. * [QP Kogo2] [QP kto1] e celunal?

122

As previously argued, the derivation of such questions contains multiple QPs, which predicts Superiority effects. Consider a stage of derivation, when v-to-Asp verb movement extends the phasal properties of vP to AspP, illustrated in (61):

(61) CP C TP [iQ] T v/AspP QP1 v/AspP [uQ] QP2 v/AspP [uQ] v/Asp vP tQP1 vP tQP2 vP tv VP PHASE DOMAIN

In (61), before the edge of vP becomes opaque by turning into the domain of AspP, all QPs raise to the edge of the higher phase, otherwise their uninterpretable Q-feature will never get checked and the derivation will crash. Recall that under the unidirectional feature-checking mechanism, adopted in this dissertation, it is the feature [uQ] on QPs what forces these elements to move out of vP to the next phase edge, given that their formal imperfection cannot be valued within vP. It follows then that in Slavic the edge of v/AspP acts as an escape hatch for QPs from the phase domain, along their successive cyclic movement to CP. Finally, once QPs reach the edge of v/Asp head, vP is transferred. The reader is invited to observe that the derivation in (61) predicts the obligatory movement of QPs to the preverbal position without lookahead: movement of QPs into the edge of AspP is triggered by their own requirement

and not by C.14

14 On the other hand, as pointed out by Ángel Gallego through personal communication, the

lookahead problem is no longer relevant at the point when movement to AspP takes place: under current Chomsky’s (2007, 2008a) phase-head-driven version of Phase Theory, C and T (and hence Asp, as a part of the extended T) are introduced at the same time.

123

Let us now consider Russian multiple wh-questions, which, as I argue, contain a single QP:

(62) a. [QP Kto1] [XP kogo2] uvidel? (Russian)

who.NOM who.ACC saw

‘Who saw whom?’

b. [QP Kogo2] [XP kto1] uvidel?

who.ACC who.NOM saw

Recall our suggestion that in such questions one Q (Q1) is merged with a

wh-containing XP in its argument position, projecting a QP, while another Q (Q2)

is merged with the interrogative C. As expected, such C head becomes specified

with two features, [iQ] and [uwh], the latter being contributed by Q2 (see section

3 for discussion). Our phase-based theory makes clear predictions regarding partial wh-fronting of the XP which is not dominated by any QP. Consider (63), representing the derivation at the stage when v-to-Asp head movement extends the vP phase domain to AspP:

(63) v-to-Asp movement extends phasehood of vP to AspP (Russian)

[CP Q2+C[iQ][uWH] [TPT [v/AspP v/Asp[vP QP[uQ] XP[iWH] v [VP … ]]]]]] Transfer

As a consequence of phase extension and PIC, the edge of vP becomes opaque to further operations. Therefore, the single QP is forced to move to the edge of AspP, in order to be able to check its [uQ] at some higher position. However, the XP, which is not dominated by any QP, does not bear any uninterpretable feature. So what does force movement of XP to the edge of AspP? I suggest that XP escapes the domain of the extended phase, before it becomes opaque, in order to be visible to C and to check its uninterpretable instance of the wh-feature in an upward manner. This is represented in (64a). If XP remains in-situ, the derivation will crash, being wh-agreement sensitive to locality, (64b):

(64) All potential goals (QP & XP) escape the domain of AspP (Russian)

a. [CP Q2+C[iQ][uWH] [TPT [v/AspP QP1[uQ] XP2[iWH] v/Asp [vP t1 t2 v [VP ]]]]]]

Transfer

b. * [CP Q2+C[iQ][uWH] [TPT [v/AspP QP1[uQ] v/Asp [vP t1 XP[iWH] v [VP ]]]]]]

Transfer

124

Under the analysis in (64a), the fronting of the second wh-phrase in a multiple question in Russian follows straightforwardly. Moreover, the view also accounts for an interesting piece of data regarding Russian single wh-questions, discussed in the previous chapter. Recall that in Russian single wh-questions the wh-phrase is not required to move necessarily to the leftmost position in the clause. Rather, sometimes the fronted wh-phrase can follow the subject (presumably, hosted at Spec,TP), (65a), or appear between the auxiliary verb (in T) and the infinitive (in v), (65b):

(65) a. Petya komui pozvonil ti? (Russian)

Peter.NOM who.DAT phoned

b. Maša budet komui znovit’ ti?

Masha will who.DAT phone.INF

‘Whom will Masha phone?’

Single questions as in (65) are quite productive in Russian. However, they can be hardly captured by any theory which assumes obligatory wh-movement to CP. However, if we assume that the wh-phrases in (65) are contained within an XP and the Q-particle is merged directly at C, the corresponding derivation proceeds straightforwardly:

(66) Derivation of single questions with partial wh-fronting (Russian)

[CP Q+C[iQ][uWH] [TPT [v/AspP XPi[iWH] v/Asp [vP ti v [VP ]]]]]]

Transfer

The Q-particle, being merged with C, contributes the uninterpretable wh-feature to the interrogative head in Russian. This feature can be checked through the upward wh-agreement in the local domain between C and the matching goal, the

wh-containing XP. In order for this agreement to take place, the XP must raise to

the edge of the phase AspP, from where it is visible to the wh-probe on C. Moreover, since the uninterpretable instance of the Q-feature on the interrogative head is checked by the Q-particle itself, no wh-movement to CP is required and the wh-word remains low.

Another piece of evidence in favor of partial wh-fronting to AspP, an operation which is forced by PIC, comes from behavior of wh-phrases in Russian

yes/no questions. For instance, observe from the following examples that the

wh-phrases kogo ‘who.ACC’ in (67) can be interpreted as a non-interrogative

125

(67) a. Ty kogo videl? (Russian)

you who.ACC saw

‘Did you see anybody?’ b. Ty videl kogo?

you saw who.ACC

c. #Kogo ty videl?15

who.ACC you saw

(68) a. Ty kogo-nibud’ videl? (Russian)

you someone.ACC saw

‘Did you see anybody?’ b. Ty videl kogo-nibud’?

you saw someone.aCC

c. * Kogo-nibud’ ty videl?

someone.ACC you saw

Observe that the wh-phrase in (67) is interpreted not as an interrogative

wh-pronoun, but rather as a wh-indefinite (cf. (68)) (for discussion, see section 3.2

of chapter 2). The (a) and (b) sentences in (67) and (68) are interpreted as

yes/no questions (‘Did you see anybody?’), rather than wh-questions (‘Who did you see?’). They are characterized by a representative raising intonation, the

interrogative stress on the verb (the wh-constituent being stressless).

I propose that the (a) sentences in (67) and (68), yielding to a SOV order, are derived along similar lines, such as sketched in (66), as shown below:

(69) Derivation of yes/no questions with partial wh-fronting (Russian)

[CP Q+C[iQ][uWH] [TPT [v/AspP XPi[iWH] v/Asp [vP ti v [VP ]]]]]]

Transfer

Namely, the Q-particle is merged high, within CP. The wh-containing XP escapes the phase domain by raising to Spec,AspP and becomes visible to C, which checks its formal imperfection. However, by not being dominated by any QP, the

wh-element does not have reasons to keep moving into CP. In fact, the

ungrammaticality of the (c) sentences in (67) and (68) confirms this observation. I tentatively suggest that the wh-in-situ option, as in the sentences (b) in (67) and (68), is legitimized in yes/no questions, because the Q-particle merged

15 Observe that the question in (67c) cannot be interpreted as a yes/no question, as opposed to

(67a,b). With the wh-phrase at the leftmost position of the sentence, the only available interpretation is that of a wh-question.

126

within CP does not always bear the wh-feature. Thus, C is not endowed with any formal imperfection and does not require local wh-agreement with any goal. Consequently, the wh-item can remain in-situ. This alternative derivation is represented below:

(70) Derivation of yes/no questions with wh-in-situ (Russian)

[CP Q+C[iQ] [TPT [v/AspP v/Asp [vP XP[iWH] v [VP ]]]]]]

Transfer

Moreover, the structures in (66) and (69) are consistent with an existent observation in the literature that there is a strong correlation between verb movement and the possibility of moving out other vP-internal material. For instance, the well-known phenomenon of the so-called Object Shift —leftward movement of a definite object across adverbials— is considered to be parasitic on verb movement (see Holmberg 1986, 1999, 2000, 2005; Bobaljik & Jonas 1996; Chomsky 1993, 2001; Svenonius 2000; Fox & Pesetsky 2005; den Dikken 2007; Gallego 2007, 2010, and many others).

To sum up this section, I proposed that Q/wh-movement in MWF languages proceeds through the edge of v/AspP, a projection which has a phasal status in Russian. Since this position is housed inside the local domain of the interrogative C, the latter can enter into wh-agreement with a wh-phrase, which is previously fronted to the edge of the extended phase. This view naturally captures the striking fact that in Russian multiple questions with a single QP, all wh-items must undergo fronting above the verb. Under this view, both QP and XP undergo movement to the multiple specifiers of AspP. The latter stays at the edge of the new phase, but the former must undergo further movement in order to check its formal imperfection, [uQ]. I address this last step in the next section.

Documento similar