Complement Anaphora in Spanish: Proportional References and Discourse Relations

Abstract

Although still disputed by some authors, complement anaphora interpretations are widely accepted today. However, they are marked and subject to strict interpretation conditions. Most commonly, monotone decreasing quantifiers facilitate a salient complement set interpretation that is suitable for subsequent anaphoric reference, whereas monotone increasing quantifiers appear to block such interpretation. In this paper, I investigate the possibilities of anaphoric reference to complement sets in Spanish quantified expressions of type QUANTIFIER(A)(B) and argue that Spanish aligns with the general pattern observed cross-linguistically according to which anaphoric reference to the reference set (refset) is default regardless of monotonicity type, and that anaphoric reference to the complement set (compset) is only possible with monotone decreasing quantifiers. My claim is that the observed discrepancy can be explained in terms of the different proportions denoted by the quantifiers, and propose an explanation based on discourse relations. My approach provides additional cross-linguistic evidence in favor of the general observed pattern, and a promising path for future investigation into the refset/compset divide.

Anaphoric Reference to Quantified Expressions in Swedish

“…(1996) had 21% refset continuations following negative QEs. Our results show quite clearly that refset is not the default interpretation for negative QEs in Swedish and, given the low ratings it is even questionable whether it is always possible, as claimed for English (Nouwen 2003) and Spanish (Zulaica-Hernández 2018). It should be noted, however, that all three studies use different methods.…”

“…In an introspective study of Spanish, Zulaica-Hernández (2018) confirms that the general pattern identified for English holds also for Spanish: positive quantifiers can only be followed by anaphors referring back to the refset , while for negative quantifiers anaphoric reference to either the refset or the compset is possible. According to Zulaica-Hernández , two principles are at play.…”

“…Moxey and Sanford 1987; Sanford et al. 1996) and Spanish (Zulaica-Hernández 2018). The finding that refset continuations were dispreferred to such extent following negative QEs is surprising though.…”

“…As mentioned above, the positive QE några (‘some’) and the negative QEs få (‘few’) and inte alla (‘not all’) stood out within their groups, as they gave rise to less uniform judgements than the other QEs. One might ask if at least part of the variation could be explained in terms of set size, as proposed by Zulaica-Hernández (2018). Looking at Spanish, Zulaica-Hernández claims that while the refset is always the default set for all QEs, they differ in how easily they appear with a compset depending on how large proportions they pick out.…”

“…2014). In an introspective study, Zulaica-Hernández (2018) reports patterns for Spanish similar to those in English. However, he claims that a crucial factor is also the relative sizes of refset and compset counted in number of members.…”