Prepositions in Spanish: Exploitation to yet unaccounted uses

MANGIALAVORI RASIA, MARÍA EUGENIA

Amsterdam

Conferencia; Going Romance 2021; 2021

University of Amsterdam (UvA) and VU Amsterdam

Prepositions in Spanish and possible exploitation to unaccounted usesM Eugenia Mangialavori RasiaCertain prepositions like [P] hasta find a dramatically divergent use systematized across Central American [CAM] Spanish varieties. In Mexican, Colombian, Peruvian, and Ecuador Spanish, namely, hasta shows a particular behavior with copular verbs in spatial (1) (Lope Blanch 2008, Bosque & Bravo 2011) and temporal constructions (2). Most importantly, divergent use coherently extends from stative to inceptive (3), and terminative verbs (4). Its distribution, along with its natural composition with additional Advs (adelante (2)b, arriba etc.) is crucially unavailable in more conservative (e.g. Iberian) Spanish varieties [IS]. The pattern is striking also for more general (non-language specific) reasons: namely, directional (left) boundary Ps are unexpected with states, certainly not in pure locatives like (1) or punctual happenings (3). Interestingly, (2)¬¬ and (5) are possible in IS to asymmetric (before/after) readings. (1)a. La casa está hasta la puntadel pueblo/hasta la derecha (CAM) thehouseisESTAR up to the tipof-the village ?The house is at the end of the village? (*IS)b. La casa está hasta adelante, hasta el fondo de la calle. (*IS)the house isESTAR until ahead until the end of the road ?The house is ahead, at the end of the road? (2)a. La tarea estará terminada hasta el final del día.(*IS) work beESTAR.FUT finished up to the end of-the day ?The work will be finished by the end of the day? b. La inscripción es hasta este mes de junio. (*IS)the enrollment is until this month of June. ?Enrollment is (=starts) next June?. (3)El proceso se inicia/comienza hasta la enseñanza superior. (*IS)the process begins/startsuntil the teaching superior ?The process begins/starts in graduate school(4)Esto acaba (recién) hasta que se cuenten todos los votos. (*IS)thisends only up tothat se-cl countallthe votes?This ends once all votes are counted?(5)a.. La Selección estará completa hasta el lunes. [lit.: the team beESTAR.FUT complete until Monday] ⇒ ?The Team will be complete by Monday? (CAM) | ?The Team will be complete until Monday? (IS)b. La tienda abre hasta las 7 (lit.: the store opens up to 7) CAM=startpoint | IS=endpoint⇒ ?The store opens at/starting from 7? (CAM) | ?The store opens until 7? (IS)Facts: contrasts with standard P choices (e.g. estar a/en) dismiss a semantically trivial innovative use. This, and the systematicity in the patterns and verbs allowed: (i)avoids coercion solutions to the alleged conflict between a directional P and a stative verb; (ii)is not coherent with putative elision of neg operators (6); (iii)points to a distinct, richer semantic denotation allowed by hasta, independent of ?redefinitions? or multiple entries. Consistent use in left-bounded predications across CAM is key.(6)Hasta ahora tuve tiempo de escribirte / logré desocuparme.UntilnowI-had timeofwrite-DATyou I-achieved free-me ?Only now I had time to wite/I got free?Hypothesis: CAM deploys hasta to measure distance from an implied starting point marking a standing [time/space] location of the speaker. Preserving a left-boundary semantics, it is compositionally used to set the endpoint of some preparatory phase preceding eventuality onset. IS is more limited, constrained to mark the endpoint of the interval where the verb-denoted eventuality holds. Consistent distributional patterns (verbs) in CAM show that hasta coexists and contrasts with the ?standard? distribution, opening the option to operate on a distance or perspectival path to be bounded. Such a preparatory phase component has been long argued across Romance varieties (e.g. (5)a, Brucart 2012). We draw on a general condition on verb+directionalP combination (Cresswell 1978). An ENDPOINT CONDITION (8) agrees with the proposal that hasta can be associated with a predication visibly distinct to the one yielded by locative Ps, somehow benefitting from its directional nature to different but coherent results with relevant verbs. The distinct semantics of constructions like (1), if related to perspectival location, is perfectly amenable to the additional variable introduced by from here in (7)a-b (abstract path). Notably, the EPC depends on the denotational properties of the directional P, which must introduce a contextually determined point of view from which the object is situated. In CAM, this entailment is strong enough to render the adjunct redundant (9), allowing the PP only if hasta is dropped. It is omissible when the verb encodes the inception component, but not free. Without hasta, (2)a-(3)a miss the ?from here? sense of perspective from a startpoint which somehow foregrounds the interval prior to inception.(7)a. The house is behind/outside/acrossthe woods (from here)(Zwarts 2005(3))b. The car is one mile from the garage/to the east.(8)Locative Ps appear with copula (be) in locative sentences. With directional Ps this is possible if the location is understood as the endpoint of a hypothetical journey described by the P from an implicit point of view (7)a, or measure phrase (7)b. (Cresswell 1978:112, Zwarts 2005:742)(9)La casa está (hasta) detrás del lote (*?desde aquí). ?The house is behind of the lot (from here)?While the proposal follows definitions of hasta like Talmy?s (2000:254), its implementation builds on analyses where the aspectual contribution of spatial Ps are laid out in terms of Vector Space Semantics [VSS] (Winter 2001 i.a.) as underlying ontology to the analysis of PP structures. On this view, CAM shows that hasta can also be used to determine a circumstance where P marks endpoint to an (abstract) set of ordered vectors, introducing a starting point and points in between on which the direction lexically encoded by the P imposes ordering, thus defining a path (Zwarts 2005:744). To this result, two assumptions are key: (i)location and other spatial properties are represented as relative positions modeled by vectors (Zwarts & Winter 2000); (ii)paths require a fixed reference object [RO] (Zwarts 2005:283) to locate the object. Even if a RO also figures in simple locative and temporal constructions (estar en/a), when applied to directional Ps this relative function accommodates the ?from here? entailment making hasta a non-trivial choice, explaining non-standard uses under a general condition like (8). Directional Ps would represent a spatial stretch that connects the starting point and the endpoint of a trajectory (located vector) with temporal and atemporal uses of directional Ps determining that the path merely preserves linear ordering (Zwarts & Winter 2000, Zwarts & Gärdenfors 2016). The set of ordered vectors (path) may be instantiated by any fictive path: a line of sight, a walking distance, or a route for a hypothetical journey. A perspectival path introduced by hasta would have an endpoint (final vector) at the RO, with a starting point at an unspecified location set by default at the location of the speaker, yielding the ?from here? entailment that accommodates the relevant condition ((7)a), yielding (9) accordingly. The projective direction (?up to?) imposes the correct ordering on the points connecting the stationary path´s start and endpoint. That P hasta is directional and, especially, a boundary P, is key to endpoint interpretation yielding asymmetric use under EPC. DPs setting a boundary to the landmark (punta, límite in (1)) also work coherently to stress the journey sense. Thus, (1)a is computed under entailment that house stands at the endpoint of an abstract trajectory traversing the town. In IS, a located vector u determines the region where the located object LO is framed (w0=location of RO and w1=relative location of the LO, Winter 2001). CAMS allows constructions where RO and LO coincide at w1 (endpoint of u), leaving the additional (contextually-fixed starting point claimed by Cresswell) (w0) as a key variable for a nontrivial sense of distance. Whereas in standard locatives w0, w1 are in the landmark, in hasta constructions w0 is not. SUMMING UP. an abstract journey sense and the idea of a perspectival location from the viewpoint of the speaker extend coherently, explaining distinct uses/combinations. In IS and CAMS, the use of hasta is equally accommodated by the P prime (birelational function) AT-END-OF (Jackendoff 1990) in spatial and nonspatial uses. In all cases, the PP expresses a PLACE that is at the terminus of a PATH starting (from here) at the RO. This explains asymmetries, as hasta marks endpoint in different ways: in IS, it limits the length of space/time the LO holds for (?until? reading), measuring the interval at which the eventuality named by the verb holds. CAM diverges to exploit AT-END-OF to measure the span of the fictive (Jackendoff 1991) (noneventive) path locating the RO at its end. CAMS would exploit this entailment, using hasta for foregrounding, in a logical extension of the same semantic (AT-END-OF) relation. This crucially outweighs the idea of meaningless anomalities (Lope Blanch). Facts point instead to understudied strategies for grammatically realizing space/time constructions with perspectival starting point as key variables, with consequences on generalizations on how predicates can be constructed and exploited in natural languages. Future research deals with possible extension/occurrence in Romance.