Clitic Climbing and Tense Auxiliaries in Polish and Romance

The aim of this paper is to compare the behaviour of Polish and Romance pronominal clitics in tense auxiliary constructions and to account for Polish facts. First, we present the system of Polish auxiliaries, briefly comparing it to Romance. Then, we discuss clitic climbing (CC), the phenomenon well-known in Romance. We contrast Polish CC with CC in Italian and French. Finally, we present a formal analysis of Polish CC. Our analysis is coached within the framework of HPSG (Head-driven Phrase Structure Grammar, (Pollard & Sag, 1987; Pollard & Sag, 1994)), which has been also used to account for CC in Romance. We follow (Borsley, 1999) in treating Polish auxiliaries as syntactic items. However, we do not find his argumentation strong enough to motivate the adaptation of the analysis proposed for French auxiliaries, (Abeillé & Godard, 1994), to Polish. We account for CC in Polish assuming that clitics can be realised independently of non-clitic arguments. Such an approach correctly explains optionality of CC in Polish without specifying what the constituent structure of auxiliary constructions looks like. 1. Tense Auxiliaries As argued in (Borsley & Rivero, 1994), past tense, (1a), and conditional verbs, (1b), are formed with an auxiliary as well: (1) a. Ty you widział seen -eś AUX.2sg ten this film. film ‘You saw this film.’ b. Ty you widział seen -byś AUX-COND.2sg ten this film. book ‘You would see this film.’ Although (e)ś and byś in (1) look like verbal inflection, they can be detached from the verb and occur on (almost) any other preceding word,1 (2) (see (Booij & Rubach, 1987) or (Spencer, 1991, ch.9) for more examples). 1Transitive prepositions and the verbal negative marker nie ‘not’ are the exceptions discussed in (Borsley & Rivero, 1994). (2) a. Ty you -ś AUX.2sg widział seen ten this film. film ‘You saw this film.’ b. Ty you byś AUX-COND.2sg widział seen ten this film. film ‘You would see this film.’ The behaviour of‘floating inflections’ illustrated in (1) and (2), has been often discussed in the literature, e.g., (Mikoś & Moravcsik, 1986), (Booij & Rubach, 1987), (Rappaport, 1988), (Spencer, 1991), (Borsley & Rivero, 1994), (Borsley, 1999). We follow here (Borsley & Rivero, 1994) and (Borsley, 1999) and treat boldfaced forms in (1) and (2) as consisting of a weak auxiliary (a clitic) and a participle. We adopt the analysis of (Borsley, 1999) of verb forms in (1) and treat them as morphological compounds (complex verbs). We also follow (Borsley, 1999) and assume that weak auxiliaries are syntactic items. However, our analysis of auxiliary constructions, such as in (2), will be different. (Borsley, 1999) represents weak auxiliaries as (subject-raising) syntactic verbs which subcategorize for a participle, similarly to the future auxiliary. He argues, however, that the future and weak auxiliaries should have different complementations. The former takes a VP complement (a traditional hierarchical structure results, (3a)) while the latter form a complex (syntactic) predicate with the participle. Hence, weak auxiliaries subcategorize for a verbal complement (a participle) as well as its complements, which results in a flat syntactic structure, (3b) (similarly to French and Italian tense auxiliary constructions, cf. (Abeillé & Godard, 1994) and (Monachesi, 1997a), respectively).