Cross-Romance variation: alternatives given by a (same?) derivational morpheme

MANGIALAVORI RASIA, MARÍA EUGENIA

Cracovia

Conferencia; THE ELEVENTH CONFERENCE ON SYNTAX, PHONOLOGY AND LANGUAGE ANALYSIS; 2018

The Jagiellonian University

Cross-Romance variation: alternatives givenby a (same?) derivational morphemeMa. Eugenia Mangialavori Rasia-CONICETPROBLEM:Themorpheme derived from the present infinitive form of Latin *-idiō (Tronci 2015, Cockburn 2012)became part of the very core of derivational morphology, rendering a qualitative and quantitative productivealternative for verb formation across the board in major Romance languages (1), with interesting parallels inGreek ízō(Necker&Tronci 2012). Its grammaticalization, however, hasled to different resultsin each case, givingverbs with significantly distinctsemantic and syntactic properties. Notably, the role of this verbalizer in Italianposes an empirical question if pared against other Romance languages, but also (synchronically) with Latin.Unlike equivalent verbalizerslike (Port., Cat.) -ejar, (Spa.) -ear, and even its Latin predecessor -izare; Italian -eggiareis systematically used in the production ofunergative verbs with equally consistentsemanticimplications.Four observationsare key:(i)-eggiareparticipates in Italian subsystem of verbal derivation as a highly productiveverbalizer (vo) yielding denominal and deadjectival verbs with coherentstructural properties; (ii)these propertiesdefine a maximal contrastwith the transitive/unaccusative, change-of-state (COS) verb given byits ?equivalent?inother Romance languages; (iii)theoppositionmotivates in Italian a nontrivial derivational alternationproducinga significantset of minimal pairs like (2), showing amorphotactically transparent contrast between zero-suffixedfientive C(hange)-O(f)-S(tate) (?become √?)) verbs with causative alternation, and (unergative)similative essivevariants(?be/looklike √?))given by-eggiare;(iv)such an unergative-stativevariantis either missing(Spa./Br.Prt.)or morphologically indistinguishable in otherRomance languages(Cat..). Based on these facts, I argue that(A)thederivational alternative (2)defines a consistent crosscut in event and argument structure, whit(non)eventivitycorrelatingwith (a)transitivity. (B)a similar structuraloptioniscollapsed under the same morpheme(e.g. Catalan)or (C)unsystematically realized by different verbalizers in other Romance languages(e.g. Latin).(1) Ita.-eggiare; Cat./Port/Prov.-ejar; Fr. -oyer; Sp. -ear;Gal. -izar, Friul. -iar, Surs. -egiar, Cerd.-iare(A)Italian ≅ Catalan. -eggiare verbs (EVs) are distinct in that they conveya state resembling or close enoughto the property/thing named by the root (√), but with a strong implication of never fully reaching this condition(Carratta&D´Alberti 2013, Necker&Tronci 2012). This derivational alternative defines a contrast in eventstructure correlatedwith a consequent contrast in argument structure. ASPECT: zero-suffixed verbs generally showvariable telicitybearing on the type of scale associated to the lexicalized property,followinga well-known relationlong noted in deadjectivals (Hay et al. 1999 i.a.)(5)a. EVs break this pattern by yielding invariably atelic forms,independent of lexicalized scale(5)b.Atelicity is nonetheless expected, given thestative nature of these similativevariants. Importantly, however, Italian statives seem to lack the event variable noted in their stative Catalancounterparts, wherenonresultativity rather follows fromby a never-ending (approachingOltra&Catroviejo 2013)transition towards the property named by √. Catalan EVs therefore license locative/temporal, manner andprogression adverbials, showing not only mixed stative/eventive status, but also a sense of incremental scalarity(and gradability) crucially missing from Italian EVs(cf. (4)) vs.(Ita.) ?ha gialleggiato un po?(#dal primo lavato)?it has looked yellow(ish) (#since the first wash)). SYNTAX: Italian also differs by offering a set of well-known testsfor split intransitivity. Here, auxiliary distribution (6), unaccusative/reflexive morphology (7), ne-cliticization(8),passive/middle formation, indicate that lack of event progression/endpoint (9)-(10), restriction to generic tenses(cf. arrossì[√RED.PRF]vs. ??rosseggiò[√RED.egg.PRF])and oddity in progressive (10)correlate with unergativity.(3) El cel fosqueja (en el fondal de l´horitzó/gradualment/naturalment/com sempre). (Cat.)?The sky is going dark(er)(at the back of the horizon/gradually/naturally/as always)?(4) La camisa ha groguejat una mica #(des de la primera rentada).(Oltra & Castroviejo 2013) (Cat.)?The t-shirt has gone yellow #(since the first wash)(5) a. sbianchire (#per/okinun?ora) |schiarire (okper/#in un?ora)?whiten?|?lighten?(in/for an hour)? (Ita.)b. biancheggiare (*in/per un?ora) |chiareggiare (*in/per un?ora)?be white/light(ish) (for/*in an hour)?(6) {ha/*è} rosseggiato/verdeggiato/biancheggiato. Cf. {*ha/è} arrossito/inverdito/sbiancato/ingiallito.(7) La pelle {*(si) arrossa/(*si) rosseggia}. ?The skin (inch) reddens?vs.?The skin (*inch) looks red(ish)?(2) √white √flat √black √round √blue MEANING BEHAVIORsbiancare appianare annerire arrotondare azzurrare ?become√(er))? ±telic COSbiancheggiare pianeggiare nereggiare rotondeggiare azzurreggiare ?be/look {√}(ish)?atelic SIMIL(8) Anche i più impudichi ne {arrossano/*rosseggiano}.?Even the most indecent ones gored?(lit.:redden)(9) {annerire/#nereggiare} (gradualmente/completamente)?blacken/look black(ish)(gradually/completely)(10) Il vetro {nereggiava/#stava nereggiando} ?The glass looked (imp)/was looking (prg)blackish? (cf. (2)a)(B)Cross-Romance results: non-homogeneous (nor systematic). Italian generally preserves the stative-unergative statusin EVs(drawing the nontrivial contrast between transitive/unaccusativeCOS sbiancare ?whiten? and the stativeunergative biancheggiare ?look white(ish)?), Spa. and BRPort. analogues give instead inchoative/causative COS(SP blanquear/ BrPRT.alvejar?whiten?),with variable telicity according to√type, thus collapsing the contrast, notedin Italian, between -ear/ejar (-eggiare analogues) and zero-suffixed forms. In turn, Catalan speakers report anontrivial ambiguity between (a) an atelic, non-resultative unergative stative, somehow paralleling Ita.biancheggiare, and (b) a transitive/unaccusative fientive COS variant amenable to Spa. blanquear/BRPrt. alvejar(11). The alternative inevent and argument structure realization accommodatesotherwise puzzlingbehavior likecompatibility with ne/se-cl.(reported in Oltra & Castroviejo 2013)and the unavailability of unergative reading inperfective tenses (13). If correct, this means that Catalan retains the event/argument structure crosscut andalternation which is derivationally nontrivial in Italian (cf. (11)vs.(2)). French (14) and EUPortuguese allow thestative atelic alternative, which is, however, visibly restricted in productivity (vs. Italian (2))(11) a.La camisa blanqueja.?The shirt lookswhite(ish)? (ATELIC(UNERGATIVE)SIMILATIVE)b. Ha blanquejat la camisa/La camisa s´ha blanquejat. ?[he] whitened the shirt/the shirt whitened?(COS)(12) Estava blanquejant *(diners negre)?(He) was laundering (lit. ?[he] whitening) black money?(13) Ha blanquejat tota la paret→?[he] whitened the whole wall?/*?The whole wall looked white(ish)?(14) blanchir/blanchoyer, rougir/rougeoyer, verdir/verdoyer, noircir/*noiroyer, bleuir/*bleuoyer (Fr.)(C)Systematicity does not trace back to Latin:Zero forms yield unergatives amenable to adjectival predication, whileinchoative variants are realized via the morpheme -esc (15). Also here, atelicity correlates with similativity(Thompson 2012:119). Importantly also, these verbs entail some sort of nonresultative COS (Camps 1969)described as mere ?intensification? of the property (reminiscent ofCatalan D-statesgiven by -eggiare?sequivalent-ejar). Notably, bounded-scale-rooted albescere (√WHITE-esc-INF)?start growing whit(er)? is just as atelic andnonresultative asCat. groguejar (√YELLOW-ej-INF)?go yellow(er)? or blanquejar(√WHITE-ej-INF) ?go white(r)?(Oltra&Castroviejo`s gloss)in itsstative similative variant (vs.the morphologically-identical COS variant(11)b).Distributionin Latin, however, is farfrom the relative systematicity and productivityseen in Italian. First, studiesshow stative uses paralleling -eggiare (Clarke 1998:13) and significant lexical gaps in the paradigm (15) (e.g.okacuo ?sharpen? /*acuesco/*exacuo ?sharpen up?). Second, some zero forms render the alternative trivial byyielding inchoative transitive (Haverling 2000) forms with variable telicity (e.g. obscuro ?darken (up)?), while -sco forms are often found in resultative uses (Gaffiot 1934:94, Dilke 1967). Even if zero forms are productivelyused for variable telicity COS, with ex-/ad-/in-prefixes associated to resultativity (Mateu 2017), the similativity of-sco can override this effect, giving a sort of scalar property increase in which nonresultativity and atelicityapparently follow from an inchoative sort of COS in which reaching the endstate defined by (the property denotedby) √ is not part of the denotation of the verb (e.g.Lat.inalbesco ?to begin to grow white? Adam 2015:2012; cf.Cat. blanquejar?go white(ish)?, where atelicity/nonresultativity rather follow from a never-ending ?approaching?relation to √). Finally, the asystematicity of -idiare/izare (the Latin analogue of -eggiare), visibly restricted for?deadjectival? formation, renders a set of denominals (Cockburn 2012) including: similative stative/behaviorverbs (betizo (√beet-IDIARE)?be/look like a beet?) loosely resembling Italian similative statives; instrumental or?manner-incorporation? verbs (LATspongizo ?clean with a sponge?); telic location/locatum verbs(sinapizo(√mustard-IDIARE)?cover with mustard?) and transitive COS. The results into a significant variabilitycontrasting with the more systematic Italian result (cf. the crosslanguage contrast in (17) vs. the systematicitybetween(Lat) graecissareand (Ita)grecheggiare, both meaning?haveGrecian manner or tone?). A nontrivial setof equally asystematic -izare/idiare verbs (Cockburn 2012:168) showsthat in Latin the split between causativeCOS and stative similatives is derivationally trivial, as opposed to Italian (2).(15) a.albeo,rubeo, vireo, flaveo, liveo ?be {white/red/green/yellow/blue}? →atelic NONSIMILATIVEDSTATEb. albesco, rubesco, viresco, flavesco, livesco?grow (white/red/green/yellow/blue)er? →atelic SIMILATIVECOS(16) candeo, albeo?be light/white?|candesco/excandesco, albesco/exalbesco ?growlight(er)/whit(er)?(17) (Lat.) latinizare ?turn/translate into Latin?|(Ita.)latineggiare, ?resemble a Latin? ?talk using Latinisms?