Scales, less-known scalar bounds and contrasting telicity

MANGIALAVORI RASIA, MARÍA EUGENIA

BERLIN

Workshop; ENDPOINTS AND SCALES; 2018

Humboldt Universitat

Less-known scalar bounds, scalarity and telicity across deadjectival verbs.Maria E Mangialavori Rasia (CONICET)Scales, particularly if conceived as abstract representations of measurement (Kennedy 2009 i.a.), have been fruitfully employed in the analysis of aspectual properties in verbs and adjectives. Previous semantic work (e.g. Hay et al. 1999) established a positive relation between scalar functions associated to gradable properties and Aktionsart. This general correlation explains major empirical patterns and allows the differentiation of natural aspectual classes (Hay 1998, Kennedy & McNally 2005 i.a.). In particular, the correlation between aspectual boundedness in adjectives (1), and verbal telicity in deadjectival verbs [DV] (2), received increasing attention in the last decades (Bertinetto & Squartini 1995, Jackendoff 1996, Ramchand 1997, Rotstein &Winter 2004), and has lent support to the claim that scalar boundedness is a significant component of the lexical aspect of verbs and adjectives (Levin & Rappaport 1995, Kennedy & McNally 2008). A related, specific claim considered here is that telicity in DVs is a non-configurational result dictated by inherent Root boundedness (as a [±bounded] scalar feature coded in √, e.g. Harley 2005).(1) a. (completely) straight, empty, dry, clean ⇢ CLOSED-SCALE ADJECTIVESb. (#completely) long, wide, soft, short ⇢ OPEN-SCALE ADJECTIVES (Hay et al. 1999)(2) a. straighten, empty, dry, clean (completely/in a minute) ⇢ BOUNDED SCALAR ROOT>TELICb. lengthen, widen, soften, shorten (#completely/#in a minute) ⇢NONBOUNDED SCALAR ROOT>ATELICMy goal here is to test the degree of generality of the scale/Aktionsart correlation in Romance DVs, based on three specific questions: (A) Are other types of scalar bound (non min./max.values) possible? Are they criterial to telicity? (B) Are derivational variables relevant? Do other DV types show the same (Property-Scale-to-Event [PSEH]) homomorphism? (C) Is argument structure correlated? These questions concern three different potential domains of variation. My aim is to identify factors systematically facilitating or constraining the homomorphism. Based on the results, I advance a more constrained account, where the lexically-supplied scale appears as a relevant locus of aspect determination only under certain conditions.(A) Non-configurational variation. Major patterns show that Italian DVs preserve the homomorphism noted in English (Harley 2005:53, Hay et al. 1999:41, also Krifka 1992, Wechsler 2005, Kennedy 2007) between the scalar function supplied by the property-denoting lexical source and verbal telicity (cf. (2) vs (3)-(4)).(3) a.[-b] (s)chiarire, scurire, incupire, infoscare, impallidire (#completamente/#in un?ora).lighten darkenv gloomv dim palev completely / in an hourb. [+b] illimpidire, opacare, arrossare, annerire, sbiancare, inverdire, ingrigire (completamente/in un?ora).clear [out] opaquev redden blacken whiten turn green turn grey completely / in an hour(4) a. Il vetro si stava [-b]schiarendo/[-b]scurendo. ⇒ Il vetro si è schiarito / scurito.?The glass was growing lighter/darker? (lit. lightening/darkening) ?The glass lightened/darkened?b. Il vetro si stava [+b]sbiancando/[+b]annerendo. ⇏ Il vetro si è sbiancato/annerito.?The glasses were turning white/black? ?The glasses turned white/black?A study of PSEH using hue DVs as toy example is theoretically and empirically motivated, especially since color forms have been associated to a less-explored type of scalar boundary contrasting significantly with the two types (lower/upper scalar ends) commonly analyzed. The Italian verbs at hand thereby offer a new empirical basis to explore the participation of less-known types of scalar bound (medium bound, McNally 2011, Hansen&Chemla 2015) in various aspectual properties (culminativity, resultativity), but also with a view to refining commonly accepted notions on scalarity. Namely, given (3)(4), color roots allow us to test the idea that scale boundaries need not be maximal or minimal elements on a scale to dictate telicity (e.g. Bertinetto & Squartini 1995, Kearns 2011). Medium scalar bounds also accommodate patterns escaping coarser-grained typologies, which predict that scalar change on bounded scales is incompatible with partial progression (5). Moreover, by handling resultativity independent of further progression, color forms point to a nontrivial scalar gap between the (medium) value yielding the expected ?result state?, and maximal progression, hence raising a question on the notion of endpoint involved in calculating telicity,culmination, resultativity. Color DVs would thus offer empirical evidence allowing to differentiate event endpoint (yielding the expected ?result state?) from scalar endpoints (max. progression allowed by a scale).(5) Si annerì, ma non completamente ?It blackened, but not completely?; appena sbiancato ?barely whitened?(B) Non-lexical variation. Alternative derivations available in Romance, giving nontrivial minimal pairs from the same roots (6), support the claim that PSEH is not general/default for DVs, inasmuch as other DV types display consistent atelicity/telicity independent of root-encoded scalar boundedness. The cases at the table show that specific conditions placed by the verbalizer are crucial to scalarity, insofar as -izzare yields telic transitive COS, while -eggiare consistently produces atelic unergatives on the same √. Argument structure tests (auxiliary distribution (7), unaccusative/reflexive cliticization (9), ne-cliticization (8)) converge with event type tests (no fact anaphor, lack of habitual reading (9), progressive (10), state reading with measure (11)) to show that in -eggiare verbs unergativity correlates with processlessness. The eventivity of -izzare DVs is key to avoid unwanted conclusions: e.g. that PSEH holds for dynamic DVs, as opposed to stative (-eggiare) alternates; or that eventiveness is a sufficient condition for PSEH (rather than a necessary one).(6) ITALIAN SPANISHa. schiarire aclarar ⇢ VARIABLE TELICITY: schiarire (in/per un?ora) ?lighten (in/for an hour)?b. chiarificare clarificar ⇢ INHERENT TELICITY: chiarificare (in/#per un?ora) ?clarify #for an hour?c. chiareggiare clarear ⇢ INHERENT ATELICITY: chiareggiare (#in/per un?ora) ?be light(ish) #in an hour?(7) {ha/*è} rosseggiato, verdeggiato, biancheggiato. Cf. {*ha/è} arrossito, inverdito, sbiancato, ingiallito.(8) Anche i più impudichi ne {arrossano/*rosseggiano}.?Even the most indecent ones go red? (lit. redden)(9) Fiori che {biancheggiano/*(si) sbiancano al sole}. ?Flowers that look white(ish)/go white in the sun?(10) Il vetro {nereggia/*stava nereggiando} ?(intended)The glass was looking blackish? (cf. annerire in (4)b)(11) Relevant contrast a. La pelle rosseggia un po? | b. La pelle (si) arrossa un po?STATIVE READING (property degree) ⇢ The skin looks a little red(ish) |⇢ The skin becomes slightly red(er)EVENTIVE READING (event runtime) ⇢ ! |⇢The skin grows red(der) for a while(C) Cross-Romance (argument structure) variation. While Italian generally preserves a core stative-unergative status in biancheggiare ?look white(ish)??contrasting nontrivially with eventive-transitive/unaccusative sbiancare ?become white??Spanish and BrPortuguese derivates give telic COS verbs allowing causative alternation (Sp. blanquear/ BRPrt. alvejar ?whiten?), thus collapsing the contrast, noted in Italian, between -ear/ejar (-eggiare morph. equivalents) and zero/en-suffixed forms. In turn, Catalan speakers report a nontrivial ambiguity (12) between: (a) an atelic, non-resultative unergative form similar to Italian biancheggiare; and (b) a telic transitive/unaccusative COS form amenable to Sp. blanquear/BRPrt. alvejar. The alternative in argument structure realization in verbs like blanquejar and rossejar explains otherwise unexpected syn/sem patterns like partitive/inch. clitics (Oltra & Castroviejo 2013), middle formation, non-stative behavior, which seem crucially restricted to unaccusative/transitive frames: cf. relevance of the object for progressive (La neu està blanquejant ?*(les teulades i exteriors de les poblacions) ?The snow is whitening rooftops and exteriors of the towns?), and the unavailability of unergative reading with perfective (L´arròs pot ser de qualsevol var ietat, però els seus grans no *(s)´han blanquejat. ?The rice can be of any sort, but its grains did not (get) whiten(ed)). Catalan variability would hence preserve the event/argument structure correlation and the consequent crosscut, which is clearly not limited to ±telicity.(12) a. La camisa blanqueja. ?The shirt looks white(ish)? (UNERGATIVE-STATIVE)b. Ha blanquejat la camisa/La camisa s´ha blanquejat. ?[he] whitened the shirt/the shirt whitened? (COS)Recap. Eventiveness appears as necessary conditions for PSEH, but it does not exhaust the problem. Data provided here suggests that: (A) to dictate non-configurational (scale-based) telicity (hence, PSEH), √ must be in a relevant (measure-out) position (affectedness remaining a key variable for scalar change); yet, (B) choice of verbalizer and (C) argument structure are primary (non-lexical) constraints. I leave these corollaries open for further work. It is nonetheless clear that derivational alternatives and argument structure must be seriously taken into account in working out generalizations on deadjectival verbs, telicity and scalarity; specially as a class of unergative deadjectival verbs is a fact yet undiscussed in the literature.