Pseudo-partitive constructions of the type DP1 of DP2 often display agreement alternations when functioning as the subject of a clause, with the verb agreeing either with DP1 (High Agreement) or with DP2 (Low Agreement). Previous studies have shown that this variation is not random, but the relative contribution of syntactic and semantic factors remains controversial. This paper investigates agreement choices in Italian pseudo-partitives through an original production experiment involving 300 native speakers. The study examines the role of four factors: clause type, semantic type of DP1, verb type, and semantic properties of DP2 (animacy, gender, size). Statistical analyses based on Conditional Random Forests, Conditional Inference Trees, and Fisher’s exact tests reveal a hierarchy of factors favouring Low Agreement. The results show that the grammaticalization status of DP1, sentence type, and the argument structure of the predicate determine the structural conditions under which Low Agreement becomes available, while the semantic properties of DP2 determine its likelihood. In particular, size and gender emerge as stronger predictors than animacy, whose effect appears to be mediated by other dimensions of salience. Building on these findings, the paper proposes a syntax–semantics interface account in which agreement is determined by the interaction between semantically salient properties and their structural accessibility.
This article gives an Optimality Theory analysis of the resolution of vowel hiatus across word boundaries based on published descriptions of late 19th and early 20th century Moroccan Judeo-Spanish, also known as haketía. Our analysis builds upon recent accounts of hiatus resolution in contemporary non-Sephardic Spanish varieties. We argue that, instead of a markedness constraint requiring syllables to have onsets (Onset), the avoidance of vowel sequences (*Hiatus) is the driver of hiatus resolution strategies, which include (i) diphthongization of non-identical vowels via gliding of the higher vowel, (ii) coalescence of identical vowels, and (iii) deletion and coalescence of mid vowels. We formalize a novel account of mid vowel alternations. Correspondence constraints against vowel deletion refer to the markedness of vowel features and to the structure of the morpheme-initial syllable rhyme containing the target vowel. A markedness constraint penalizes adjacent round vocoids within a syllable. Our analysis explains why hiatus resolution fails to apply when the second vowel bears nuclear phrasal stress, a pattern also observed elsewhere in present-day Ibero-Romance. We extend the analysis to cover stylistic and typological variation in Northern-Central Peninsular Spanish and Chicano Spanish, while avoiding several complications of previous theoretical accounts.
Low não, a novel negative construction in Brazilian Portuguese, is infrequent in corpus analyses, raising the question of whether it represents a fully licensed but rare variant or reflects an emerging pattern whose grammatical status is not yet stable across speakers. This study evaluates speakers’ acceptability judgments and reaction times using a speeded acceptability judgment task to test predictions from generative approaches and usage-based accounts. Results from 37 participants show that low não is consistently rated lower than canonical não, yet higher than ungrammatical distractors. Reaction times are comparable to ungrammatical distractors, suggesting increased evaluation effort. The data also reveal substantial interspeaker variability across both acceptability and processing measures. The combination of intermediate acceptability, longer reaction times, and structured variability observed in this study is compatible with both generative and usage-based accounts but is more parsimoniously captured by usage-based approaches in which differences in experience and entrenchment give rise to graded variation in acceptability and processing dynamics. Overall, these results illustrate how experimental methods can provide insight into both synchronic distribution of novel constructions and the mechanisms that may underlie syntactic change, while highlighting the importance of modelling acceptability and processing as partially dissociable dimensions of linguistic evaluation.
This paper demonstrates how our understanding of indefinite object drop can inform our understanding of other constructions involving bare nouns and indefinite null arguments in Spanish. We extend the proposal of indefinite object drop developed by Martínez-García et al. (2025) to internal arguments of some intransitive constructions (specifically, unaccusative and passive se-constructions). Our findings reveal that the same lexical constraints that govern the distribution of indefinite object drop in Spanish also apply to indefinite subject drop in intransitive constructions. Consequently, we propose a unified analysis of indefinite argument drop. This contributes to a better understanding of the distribution of indefinite argument drop, and also of unaccusative and (analytic and se) passive constructions in Spanish. In particular, our analysis provides further support for analyzing passive se-constructions as being similar to unaccusatives and impersonal se-constructions, as independently claimed in Ormazabal & Romero (2024).
The present paper aims to analyze the properties of Spanish reduplicative coordination (ReCo) within a generative framework. Therefore, it intends to systematically contrast Spanish ReCo with the characterization and predictions made for English by De Vos (2005). English examples as (1.a) show Spanish counterparts as (1.b), suggesting that both languages have parallel syntactic structures involving a conjunction with a similar semantic interpretation: (1) a. (De Vos 2005: 62) My friend Jim Stoltz walks and walks and walks in the wilderness • Mi amigo Jim Stoltz camina y camina y camina en la naturaleza. The Spanish structure in (1.b) represents a case of pseudo-coordination, as De Vos claims for English, since the sentences do not describe two distinct events but only one event of walking which is longer than expected. However, we will try to demonstrate that there are significant differences between Spanish and English in regard to the semantic interpretation of ReCo and to its syntactic distribution with predicates of various Aktionsart classes. These contrasts deserve a detailed analysis, since they are also a symptom of the fact that each language has an essentially distinct syntactic structure, implying, in turn, that De Vos’ proposal is not applicable to Spanish.
