Practical Information - Generative Linguistics in the Old World
Practical Information - Generative Linguistics in the Old World
Practical Information - Generative Linguistics in the Old World
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Rais<strong>in</strong>g to Object from f<strong>in</strong>ite CPs: dual A/A-bar and MCCGabriela Alboiu & Virg<strong>in</strong>ia HillIssue. In colloquial Romanian (Rom) verbs express<strong>in</strong>g knowledge from reason<strong>in</strong>g (e.g.cunosc, ştiu ‘know’) or <strong>in</strong>ference (e.g. văd ‘see/realize’, aud ‘hear/f<strong>in</strong>d out’) allow for <strong>the</strong><strong>the</strong>matic subject of <strong>the</strong>ir embedded clause to surface ei<strong>the</strong>r <strong>in</strong> <strong>the</strong> f<strong>in</strong>ite <strong>in</strong>dicativecomplement CP, with NOM spell-out, see (1a), or, <strong>in</strong> <strong>the</strong> matrix clause, with ACC spell-out,see (1b). Both (1a) and (1b) have evidential read<strong>in</strong>gs.(1) a. Am văzut [că Ion k /el k e k pompier / lăcomeşte k la mâncare].AUX.1 seen that Ion/ he.NOM is firefighter / is.greedy at food‘I/We saw (= realized) that Ion is a fire-fighter / greedy with food.’b. L k -am văzut pe Ion k [că (*e k pompier) / lăcomeşte k la mâncare].CL.3SGM.ACC-AUX.1 seen PRT Ion that (is firefighter) / is.greedy.3SG at food‘I/We saw Ion be<strong>in</strong>g (*a firefighter) / greedy with food.’The evidential nature of perception verbs is not surpris<strong>in</strong>g, but <strong>the</strong> follow<strong>in</strong>g facts might be:(i) Subject rais<strong>in</strong>g changes evidentiality: <strong>in</strong> (1a), <strong>the</strong>re is <strong>in</strong>ference of a fact, while <strong>in</strong> (1b), <strong>the</strong>raised subject is evaluated by <strong>the</strong> speaker, thus rul<strong>in</strong>g out <strong>in</strong>dividual-level predicates.Specifically, <strong>the</strong>re is a shift <strong>in</strong> speaker commitment, so a shift <strong>in</strong> ‘evidence type’ (Rooryck2001) with rais<strong>in</strong>g: ei<strong>the</strong>r from <strong>in</strong>direct to direct/attested evidentiality (<strong>in</strong> <strong>the</strong> sense of Willett1988), or with<strong>in</strong> <strong>in</strong>direct evidentiality, from hearsay/reportative to <strong>in</strong>ferential. (ii) While Suto-Surais<strong>in</strong>g is known to trigger evidential mean<strong>in</strong>gs (Ruwet 1972, Rooryck 2001), Su-to-Obj rais<strong>in</strong>g has not thus been analysed. (iii) Su rais<strong>in</strong>g is out of a f<strong>in</strong>ite, Case-licens<strong>in</strong>g CP, so<strong>the</strong> trigger for this DP movement must be accounted for <strong>in</strong>dependently of Case requirements.Objective. We argue for <strong>the</strong> follow<strong>in</strong>g: (i) The derivation <strong>in</strong> (1b) arises from Rais<strong>in</strong>g toObject (RtoO)/ECM, across <strong>the</strong> phasal <strong>in</strong>dicative CP; (ii) RtoO <strong>in</strong> Rom is both A-bar and A-movement; (iii) The trigger for movement is an [Eval(uative)] feature grammaticized onto<strong>the</strong> <strong>in</strong>herently evidential ma<strong>in</strong> clause predicate with shifted evidentiality.Background. Formal analyses of constructions similar to (1b) show a split between: (i) across-clausal movement analysis, where <strong>the</strong> DP moves from a non-f<strong>in</strong>ite complement clauseto a matrix Case position (i.e. standard ECM), e.g. Bošković (2007), Bowers (2002), Johnson(1991); or (ii) an external Merge/proleptic construction, where <strong>the</strong> DP (or associated clitic) isbase-generated <strong>in</strong> <strong>the</strong> matrix clause for discourse requirements, and is cha<strong>in</strong> related to an A orA-bar position <strong>in</strong> <strong>the</strong> complement clause which, cross-l<strong>in</strong>guistically, could be f<strong>in</strong>ite or nonf<strong>in</strong>ite(e.g. Bruen<strong>in</strong>g 2001, Davies 2005, Massam 1985).Properties. First, matrix base-generation cannot be assumed for Rom, on several grounds:(i) Evaluative/evidential Vs disallow <strong>the</strong> CAUSE+HAVE/LOCATION analysis of ditransitives(Harley 2002) and are exclusively mono-transitive; (ii) The relevant DP disallows resumptivepronouns <strong>in</strong> <strong>the</strong> embedded clause, whereas object control constructions, which are ditransitive,allow <strong>the</strong>m: compare (2a) to control (2b); (iii) A relative clause analysis is ruledout due to lack of adjacency: see (3) with <strong>the</strong> <strong>in</strong>terven<strong>in</strong>g matrix subject <strong>in</strong>terven<strong>in</strong>g. Hence,<strong>in</strong> both (1a) & (1b), <strong>the</strong> matrix verb selects only <strong>the</strong> obligatorily <strong>in</strong>dicative CP complement.(2) a. Îl k ştiu pe Rareş k [că e (*el k ) om bun (*el k ).CL.3SG.M.ACC know.1SG PRT Rares [that is 3SG.M.NOM man good 3SG.M.NOM‘I know Rares to be a good man.’b. L k -am conv<strong>in</strong>s (pe Ion k ) [să plăteasca (el k ) lum<strong>in</strong>a].3CL.SG.M.ACC-AUX.1 conv<strong>in</strong>ced (PRT Ion) [SUBJ pay.subj.3 3SG.M.NOM light‘I/We conv<strong>in</strong>ced Ion to pay <strong>the</strong> hydro bill.’(3) b. Îl ştia pe Ion toată lumea [că era om bun].him knew DOM Ion all world.<strong>the</strong> that was man good‘Everybody knew Ion to be a good man.’