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

MANGIALAVORI RASIA, MARÍA EUGENIA

Pavia

Conferencia; The shaping of transitivity and argument structure: theoretical and empirical perspectives; 2018

Italian Ministry for Education and Research (MIUR)

Cross-Romance variation: alternatives allowed by a (same?) derivational morphemeMaria E Mangialavori Rasia ? CONICET (Argentina)A morpheme derived from the present infinitive form of Latin *-idiō (*-idiāre/izare) can be argued to undergo an evolutionary process (after the diachronic discontinuity seen in Latin, where it starts out as lexical borrowing from Greek only to achieve restricted productivity, Cockburn 2012, Tronci 2015), whereby it becomes part of the very core of Romance derivational morphology, displaying qualitative and quantitative productivity. Derivational analogues are found across the board in Romance (Ita. -eggiare; Cat./Port/ Prov. -ejar; Fr. -oyer; Sp. -ear; Gal. -izar, Friul. -iar, Surs. -egiar, Cerd.-iare) with parallels in Greek (-ízō, Necker & Tronci 2012). The process, however, can be argued to have led to different results, producing verbs with sufficiently variable semantic and syntactic properties across Romance. In particular, the role of this morpheme in Italian verb formation poses an empirical question if compared against the more general picture of main Romance languages, but also (synchronically) with respect to Latin. The specific question concerns the fact that, unlike equivalent morphemes like -ejar (Portuguese, Catalan), -ear (Spanish), and even its Latin predecessor -idiare/-izare; -eggiare is generally dedicated to the formation of unergative atelic verbs with nontrivial semantic implications. The primary assumptions at work here are that: (i) -eggiare participates in Italian subsystem of (deadjectival/denominal) verbal derivation as productive verbalizer (vo) drawing consistent and non-trivial structural properties; (ii) these fundamental properties are amenable to one of two forms collapsed in Latin and other Romance languages (unergative similative essive (?act/look like √?) vs. unaccusative/transitive fientive (?become/turn into √?); (iii) this contrast motivates in Italian a nontrivial derivational alternation?producing minimal pairs like (1)?i.e., zero-suffixed forms vs. morpho-tactically transparent forms with -eggiare?missing in languages where the morpheme is not committed to similativity/unergativity. Here, we will see (A) that the derivational alternative defines a consistent crosscut in meaning, event type and argument structure; but also that a similar split is (B) realized via different morphemes (Latin) or (C) morphosyntactically trivial in other Romance varieties..(1)√white | √flat| √black| √round| √blue±telic COSa. sbiancare | appianare| annerire| arrotondare | azzurrare?become √(er)? atelic SIMILATIVEb. biancheggiare| pianeggiare| nereggiare| rotondeggiare| azzurreggiare ?be/look{√}(ish)?(A) Italian variation. -eggiare verbs (EVs) are productively formed to convey a state which resembles or is close enough to the property/thing named by the root, but with the strong implication of never reaching such condition (notated as (ish) above) (Carratta & D'Alberti 2013, Necker & Tronci 2012). Previous work: the derivational alternative (1) defines an aspectual contrast correlated with argument structure. ● While non-suffixed verbs show variable telicity depending on the scalarity of the base property, according to a well-known relation in deadjectivals (Hay et al. 1999 a.m.o.); -ificare and -izzare (historically related to -eggiare) forms break this pattern by producing invariably telic, resultative verbs; in turn, EVs are invariably atelic (independent of root boundedness). While atelicity is expected, inasmuch as stativity is involved, the similative entailment (also in French, Catalan, cf.(C)) crucially correlates with lack of property ascription. (2) also shows that telicity is nonconfigurational iff √ is in the correct position (cf. [VP[Vo,√P [DP,√]]]→±TELICITY((2)a); [vP[√P[DP,√], vo]] ((2)b-c)) ● Standard tests (aux. distribution (3), unaccusative/refl. cliticization (4), ne-cliticization (5)) converge with event type tests (lack of habitual reading, scalar reading with a little bit, oddity in perfective/ progressive) to show that in EVs unergativity consistently correlates with atelicity and stativity (6)-(7). (2) a. √LIGHT.3S schiarire (Italian) aclarar (Spanish)⇢ VARIABLE TELICITY: schiarire (in/per un?ora) ?lighten (in/for an hour)?b. √LIGHT.IFY.3Schiarificareclarificar ⇢ TELIC, RESULTATIVE: chiarificare (in/#per un?ora) ?clarify #for an hour?c. √LIGHT.EGGIARE.3S chiareggiareclarear ⇢ ATELIC, NONRESULTIVE: chiareggiare (#in/per un?ora) ?be light(ish) #in an hour?(3) a. {ha/*è} rosseggiato /verdeggiato/biancheggiato. have is√RED.EGGIARE.PST√GREEN.EGGIARE.PST√WHITE.EGGIARE.PSTb.{*ha/è} arrossito/inverdito/sbiancato have is√RED.PST√GREEN. PST√WHITE. PST.(4) La pelle {(si) arrossa /(*si) rosseggia}.the skinINCH√RED.3SINCH √RED.EGGIARE.3S?The skin reddens?(5)Anche i più impudichi ne {arrossano/*rosseggiano}.also the more indecentPART√RED.3P√RED.EGGIARE.3P?Even the most indecent ones go red? (lit. redden)(6) {annerire/#nereggiare} (completamente/gradualmente) √BLACK.INF√BLACK.EGGIARE.INF completelygradually?blacken/look black(ish) (completely/gradually)?(7) Il vetro { nereggiava/#stava nereggiando}. theglass√BLACK.EGGIARE.IMP.3SBE.IMP.3S.√BLACK.EGGIARE.GER?(intended)The glass was looking blackish? (cf. annerire in (1) a)? (B)Latin Variations. Similativity independent of stativity. The alternative takes a different form in Latin. While zero derivatives yield unergatives amenable to adjectival predication (thus contrasting with (1)b), an inchoative variant is realized via the morpheme -esc (8) which introduces a similative flavor not seen in Greek (Thompson 2012:119). Such verbs are, however, atelic/non-resultative (Camps 1969). Again, atelicity correlates with similativity, which in this case yields a nonresultative COS described as mere ?intensification? of the property (hence, closed-scale-rooted albesco is as atelic as open-scale acesco ?start growing sou(e)r?). However, distribution is far from systematicity: -sco is not committed to COS, further studies show stative uses paralleling -eggiare (Clarke 1998:13); and the alternation is not fully productive (9). Moreover, some zero forms render the alternative trivial by yielding inchoative transitive predicates with variable telicity themselves (e.g. obscuro) while -sco forms are often found in resultative uses (Gaffiot 1934:94, Dilke 1967). Further, zero forms are productively used for variable telicity COS (Haverling 2000), while ex-/ad-/in-prefixes are associated to resultative predicates (Mateu 2017). Prefixation does not guarantee resultativity, however: -esco may override this effect (cf. inalbesco ?to begin to grow white? Adam 2015:2012). In turn, -idiare (the Latin analogue of -eggiare) yields various sorts of denominals (10) (Cockburn 2012) including: similative statives and behavior verbs (betizo(√beet-IDIARE)?be/look like a beet?); instrumental or ?manner-incorporation? verbs (LATspongizo ?clean with a sponge?); and inherently telic location/ locatum verbs?which produces important contrasts with both other outcomes in Latin ((10)a) and the more systematic Italian result: cf.((10)b with √graec ?Greek?, which gives LAT.graecissare and ITAgrecheggiare, both meaning ?behave as a Greek?. (10)b is also relevant insofar as the split between inchoative/causative COS (cf. eunuchizo ?turn into a eunuch?) and stative similatives paralleling ((1)b) (amethystizo ?be of the color of the amethyst?, Cockburn 2012:168) is derivationally trivial in Latin.(8) a. albeo, rubeo, vireo, flaveo, liveo √WHITE.1S √RED.1S√GREEN.1S√YELLOW.1S√BLUE.1S?be {white/red/green/yellow/blue}? ⇢ atelic NONSIMILATIVE STATEb. albesco, rubesco, viresco, flavesco, livesco √WHITE.SC.1S √RED.SC.1S√GREEN.SC.1S√YELLOW.SC.1S√BLUE. SC1S?grow (white/red/green/yellow/blue)er?⇢ atelic SIMILATIVE COS(9)a.*aureo/auresco, *nigreo/nigresco, obscuro/*obscuresco√WHITE.1S /√WHITE.SC.1S√BLACK.1S /√BLACK.SC.1S √DARK.1S /√DARK.SC.1Sb. acuocf. *acuesco/exacuo √SHARP.1S √SHARP.SC.1SEX.√SHARP.1S?sharpen? ?sharpen up??sharpen up?c. candeo, albeo cf. candesco/excandesco, albesco/exalbesco √LIGHT.1S√WHITE.1S√LIGHT.SC.1SEX.√LIGHT.1S√WHITE.SC.1SEX.√WHITE.1S?be light? ?be white? ?grow light(er)/whit(er)?(10) a. aromatizo cf. sinapizo√SMELL.IDIARE.1s√MUSTARD.IDIARE.1s ?to smell of spices??cover with mustard?b. latinizare (Latin)cf. latineggiare (Italian)√LATIN.IDIARE.INF√LATIN.EGGIARE.INF?turn/translate into Latin? ?resemble a Latin? ?talk using Latinisms?(C) Cross-Romance variations. While Italian generally preserves the stative-unergative status in EVs?drawing the nontrivial contrast between transitive/unaccusative COS sbiancare ?whiten? and the similative, stative-unergative biancheggiare ?look white(ish)??, Spanish and BRPortuguese analogues produce inchoative/causative COS (SP blanquear/BrPRT.alvejar ?whiten?) with telicity dictated by √, thus collapsing the contrast, noted in Italian, between -ear/ejar (-eggiare analogues) and zero-suffixed forms. In turn, surveyed Catalan speakers report a nontrivial ambiguity between (a)the atelic, non-resultative unergative paralleling Italian biancheggiare, and (b) a transitive/unaccusative COS variant amenable to Sp. blanquear or BRPrt. alvejar (11). The alternative in argument structure realization would explain divergent behavior like eventual compatibility with ne/se-clitics (attested also by Oltra & Castroviejo 2013 under a D-state analysis), and non-stative patterns like affinity with progressive/perfective tenses (apparently restricted to unaccusative/transitive frames, according to the nontrivial realization of the object in (12), and the unavailability of unergative reading in perfective tenses (13)). If correct, Catalan would anyhow retain the event/argument structure correlation and the consequent crosscut which is derivationally nontrivial in Italian (cf. (11)vs.(1)). French (14) and EUPortuguese allow the alternation, but productivity is visibly lower (cf. Ita. √ner/√azzurr in (1)).(11) a. La camisa blanqueja. the shirt √WHITE.EGGIARE.3S?The shirt looks white(ish)? (UNERGATIVE-STATIVE) ⇢ atelic SIMILATIVEb. Ha blanquejat la camisa./La camisa s'ha blanquejat.has√WHITE.EGGIARE.PST the shirt the shirt INCH.has√WHITE.EGGIARE.PST ?[he] whitened the shirt?/?the shirt whitened?⇢ ±telic COS(12)Estava blanquejant *(dinersnegre). was√WHITE.EGGIARE.GERmoneyblack?(He) was laundering money? (lit. ?[he] was whitening black money?)(13)Ha blanquejat tota la paret.has √WHITE.EGGIARE.PSTallthewallok?[he] whitened the whole wall?/*?The whole wall looked white(ish)?(14) blanchir/blanchoyer, rougir/rougeoyer, verdir/verdoyer, √WHITE.INF/√WHITE.EGGIARE.INF √RED.INF/√RED.EGGIARE.INF√GREEN.INF/√GREEN.EGGIARE.INF?turn white/look white(ish)??turn red/look red(ish)? ?turn green/look green(ish)?noircir/*noiroyer, bleuir/*bleuoyer√BLACK.INF/√BLACK.EGGIARE.INF √BLUE.INF/√BLUE.EGGIARE.INF?turn black/look black(ish)??turn blue/look blu(ish)? In sum. Although the general diachronic evolution of the morpheme has been addressed in detail before (Cockburn 2012, Tronci 2015), a synchronic analysis identifying potential cross-Romance variations is necessary. In Italian, the systematic properties of verbs on -eggiare arguably define a nontrivial alternative for the derivation of deadjectival verbs, at the same time that they point to key contrasts across Romance languages bearing on the semantic/syntactic role of the verbalizer in verb formation process and related phenomena, such as lexically-induced (a)telicity and argument structure.