Title: Transparency effects in Polish
Patrycja Jablonska, University of Tromsø



Transparency effects in Polish
	The aim of the presentation is to check the validity of the proposal 
made by Cinque (2000a, 2000b) concerning 'Restructuring' (Rizzi 1976). 
Rizzi argues that there exists a restructuring rule governed by a specific 
types of verbs (those include modals, aspectuals and motion verbs), 
which transforms a bisentential structure into a simple sentence. This, in 
turn, gives rise to so called 'transparency effects'. Cinque's standpoint on 
the issue is that the higher verb in restructuring contexts is a functional 
head associated with the VP projected by the lexical verb (i.e. infinitive).
When trying to establish whether transparency effects are actually 
attested in Polish I will consider the following phenomena:
q	long NP movement with sie

(1)	a. Ciezko sie czyta te artykuly.
             hard sie read-3rdsing these article-ACC
	b. Te artykuly ciezko sie czytaja.
	    these articles-NOM hard sie read-3rdpl
These two structures are analysed by Cinque as [-argumental se] ((1a)) 
and [+argumental] se ((1b)). [+arg se]  absorbs the external q-role and 
hence the verb cannot assign ACC case to its object. It should be noted, 
however, that in Polish, as opposed to Italian and Czech (cf. Medova 
2000), [+arg se] is clearly marked and subject to further restrictions. 
Let us now see how long NP movement fares in ordinary control 
structures, as opposed to restructuring contexts:
(2)	a. Przyrzeklo sie wybudowac nowe domy do grudnia 2000.  
promised3rdsing sie to-build new hoses-ACC till December 2ooo b.*Nowe 
domy przyrzekly sie wybudowac do grudnia 2000.  new housesNOM 
promised sie to-build till December 2000
(3)	modals
	a. Powinno sie dobrze sprzedac te domy.
	    should-3rdsing sie well to-sell these houses-ACC
	b. Te domy powinny sie dobrze sprzedac.
	    these housesNOM should-3rdpl sie well to-sell
(2b) is ill-formed, which indicates that long NP movement is not possible 
with subject control verbs, but it is possible with modals (3b). 
q	long passivization
On the biclausal analysis the phenomenon is puzzling, since it is not 
clear why passive morphology attached to the higher verb should affect 
Q-role assignment and Case licensing properties of the lower verb, as in 
(4):
(4)	La casa fu finita di construire il mese scorso.  this house was 
finished to build the last month
								(Cinque 2000a)
Assuming the hierarchy of functional heads proposed by Cinque, it 
follows that we don't expect to find long passives with modal verbs 
simply because Modal0 is located higher than Voice0. Still, we do hope to 
find long passives with e.g. completive verbs, because Completive aspect 
head is located lower than Voice head. This is possible in Italian, but not 
in Polish:
(5)	*Te domy byly skonczone budowac w dwa miesiace.  these houses 
were finished to-build in two months

q	clitic climbing
In Polish clitics climb in between the finite verb and the infinitive. 
Suppose that the structure is monoclausal and the finite verb left-
adjoins to the clitic head, as in (6):
(6)	Maria powinna mu je dac.
	Maria should himDAT themACCto-give
Still, the version without clitic climbing is equally well-formed, which 
might indicate that clitic climbing is optional.
(7)	Maria powinna dac mu je wczoraj.
	Maria should to-give himDAT themACCyesterday.
Now, it seems sensible to check whether long NP movement and clitic 
climbing) go together:
(8)	a. *Te domy powinny sie sprzedac im jutro.  these housesNOM 
should sie to-sell themDAT tomorrow
	b. ?Te domy powinny sie im sprzedac jutro.

Apparently long NP movement with sie cannot be separated from clitic 
climbing. I attribute the slight deviance of (9b) to the fact that the 
construction is marked and the infinitive usually does not appear with 
any object (presumably because of the generic reading). 
Alternatively, one might assume that unacceptability of (8a) is 
simply due to 'clitic split', which is illicit even under biclausal analysis 
(cf. Kayne 1994).
q	sie sie Æ sie contraction
Contraction of two reflexive clitics is optional when the lower clitic 
follows the infinitive:
(9)	 Piotrowi udalo sie nie spoznic (sie) na kolacje.
 PiotrDAT managed sie not to-be.late sie for dinner
Yet, it is obligatory if the clitic climbs up, as illustrated by 
unacceptability of (10):
(10)	*Piotrowi udalo sie sie nie spoznic na kolacje.

This is an argument in favour of monoclausal analysis, since there is 
only one reflexive clitic slot per clause. Consequently, there is no other 
option for the clitics but to amalgamate.
The present considerations also bear on the issue of universality of the 
hierarchy of functional projections as proposed by Cinque.

References:
Cinque, G. 2000a. The Interaction of Passive, Causative and 
'Restructuring' in Romance. ms, University of Venice.
Cinque, G. 2000b. 'Restructuring' and the order of aspectual and root 
modal heads. ms, University of Venice.
Medova, L. 2000. Transparency phenomena in Czech syntax. Cand. Phil. 
thesis, University of Tromsø.
Rizzi, L. 1976. Ristrutturazione. Rivista di grammatica generativa1. 1-54.