PAST PARTICIPLE AGREEMENT IN MAJORCAN CATALAN: THE RELEVANCE OF INNER ASPECT

Majorcan Catalan (at least for some speakers, especially the elder ones) admits, optionally, past participle agreement with the object in situ (PPAOIS) (1). However, a fact not observed until now is that PPAOIS is not always felicitous: it is accepted in telic dynamic events, but is very odd in stative situations and in some atelic dynamic constructions, but not in all of them. (1) S’ ha {rentades/rentat} ses mans. CL.REFL has washed.{FEM.PL/MAS.SG} the.FEM.PL hands.FEM.PL ‘{He/she} has washed {his/her} hands’. Firstly, PPAOIS seems to be infelicitous with K(imian) or pure states (2a); and also with D(avidsonian) or interval states (2b), which involve an event argument, although they still denote non-dynamic or homogeneous situations: (2) a. Na Maria sempre ha {temut/*temudes} ses bubotes. ART.PERS.FEM Maria always has feared.{MAS.SG/FEM.PL} the ghosts ‘Maria has always been afraid of ghosts’. b. Sa pobresa ha {preocupat/*preocupada} na Maria des de sempre. the poverty has worried.{MAS.SG/FEM.SG} ART Maria since always ‘Maria has always worried about poverty’. Nevertheless, PPAOIS is possible with some «stative» verbs that can be used in telic dynamic constructions (3). These are the same kind of verbs that allow accusative case in Finnish, instead of partitive case, as reported by KIPARSKY (1998) —e. g.: omistaa (‘have, own’)—, and they are called high (or level-2) pure stative verbs by JAQUE (2014), which can unfold more verbal structure —not only Init(iation), but also Proc(ess) and Res(ult), in RAMCHAND’s (2008) terms— if they are used in past simple or present perfect. (3) Has {tenguda/tengut} una idea extraordinària! have.2SG had.{FEM.SG/MAS.SG} an idea extraordinary ‘You’ve had a great idea!’.

Majorcan Catalan (at least for some speakers, especially the elder ones) admits, optionally, past participle agreement with the object in situ (PPAOIS) (1).However, a fact not observed until now is that PPAOIS is not always felicitous: it is accepted in telic dynamic events, but is very odd in stative situations and in some atelic dynamic constructions, but not in all of them.
the poverty has worried.{MAS.SG/FEM.SG} ART Maria since always 'Maria has always worried about poverty'.Nevertheless, PPAOIS is possible with some «stative» verbs that can be used in telic dynamic constructions (3).These are the same kind of verbs that allow accusative case in Finnish, instead of partitive case, as reported by KIPARSKY (1998) -e.g.: omistaa ('have, own')-, and they are called high (or level-2) pure stative verbs by JAQUE (2014), which can unfold more verbal structure -not only Init(iation), but also Proc(ess) and Res(ult), in RAMCHAND's (2008) terms-if they are used in past simple or present perfect.
Just like the Finnish accusative/partitive case distinction observed by KIPARSKY, in Majorcan Catalan we find a very interesting contrast: in (4a), with PPAOIS, all the female students the professor had are brilliant; but this is not necessarily true for (4b), without PPAOIS.( 4 'This academic-year I've had some brilliant female-students'.Secondly, PPAOIS seems also to be impossible in some atelic dynamic constructions, those ones with nouns bounded by a determiner/quantifier (5).By contrast, it is perfectly gramma-tical with bare plurals (6a) and with bare mass nouns (6b).( 5 For KAYNE (1989), PPAOIS would be a case similar to right-dislocations, with a null resumptive clitic moving through the specifier of AgrO.But for BELLETTI 2006, it would be strange that a direct object systematically be right-dislocated, and the presence of the silent clitic would need to be independently justified.I argue that KAYNE's analysis can be (at least partially) maintained, considering sentences like (7), where PPA with an NPI is ungram-matical (unless it is modified), and I propose that the pro-object moves to a LowTop position: (7) No havia presa [cap rabiada *(tan grossa)]NPI mai.not had.3SGtaken.FEM.SG any rage so big never '{He/she} had never been so enraged'.D' ALESSANDRO & ROBERTS (2008), considering the position of measure adverbs, claim that PPAOIS in Abruzzese would imply that both the participle and the object remain in a position lower than v* (a phase head), in contrast to current standard Italian (which do not allow PPAOIS).Nevertheless, as for adverbs, Majorcan Catalan behaves just like standard Catalan and Italian (8).Moreover, their proposal (and also KAYNE's) would not explain the contrasts from ( 1)-( 6), which show that, as for PPAOIS, event structure should be taken into account.( 8) a. Le so poche capite. [Abruzzese] CL.it am little understood b.L' ho capito poco.

[Standard and Majorcan Catalan]
CL.it have.I understood little 'I (have) understood it little'.I follow BORER (2005), MACDONALD (2008) and TRAVIS (2010) in assuming that a specific functional head (Asp), related to so-called inner aspect and event quantisation, can be present in the clausal structure, between vP and VP.Asp instantiates syntactically the so-called objectto-event mapping through an Agree relation, as the NP object can influence the aspectual interpretation of the whole predicate, depending on the NP being quantised or non-quantised.I propose that Asp establishes a double Agree relation with the object, in order to get two kind of unvalued features valued: on the one hand, its quantity or quantisation features (if Asp is valued as [+q], the predicate will be interpreted as telic; if it is valued as [-q], the predicate will be atelic) and, on the other hand, its [uϕ] (gender and number) features.
If Asp is not present in the structure -as in K-states, according to BORER (2005) and MACDONALD (2008)-, the impossibility of PPAOIS follows.I assume with BORER (2005), contra MACDONALD (2008), that neither do sentences like (5) have Asp, although they could involve a functional shell (F S ) or, better yet, Proc, in order to license the event argument (as argued by JAQUE 2014 to account for D-states).As a consequence, PPAOIS is also forbidden here.Thus, I argue that Asp is, in fact, Proc with [uq] and [uϕ] extra features.
Contra RAMCHAND (2008) (who claims that telicity can emerge either from the presence of a ResP or, with no Res, simply as a semantic entailment), Majorcan PPAOIS shows, like the Finnish accusative/partitive case distinction, that event quantisation is grammatically encoded.
In conclusion, in Majorcan Catalan, the morphophonological insertion of inflectional (gender and number) suffixes in the past participle would be sensitive to the presence of Asp (or Proc [uq][uϕ]), which, previously, has maintained a double Agree syntactic relation with the NP object.This is coherent with the so-called Borer-Chomsky conjecture: parametric variation is restricted to the lexicon and to a narrow category of morphological inflectional properties.