Search

The verb jau5 ‘have’ in Cantonese can be placed immediately before the verbal or adjectival predicate, forming “jau5 + predicate” construction (henceforth pre-predicate jau5), and pervious analyses consider the pre-predicate jau5 as the preposed perfective marker –zo2. The same construction “you ‘have’ + VP” renders ungrammaticality in Mandarin, with the post-verbal perfective –le being its only form. In this paper, contra previous analyses, I argue that pre-predicate jau5 and post-verbal perfective –zo2 cannot be equivalents. Following S.F. Huang's (1981) ideas on determiner quantifiers, I propose that Cantonese pre-predicate jau5 ‘have’ is an assertive existential quantifier operating in the domain of verbs and their event time, which gives rise to two assertive meanings: (i) When the situation is interpreted as non-specific, pre-predicate jau5 asserts at least one occurrence of the event; and (ii) when the situation is interpreted as specific, jau5 asserts the actualization of the natural final endpoint or the explicit bound of the situation. The existential nature of jau5 ‘have’ is inherently borne out in the lexical semantics of “have”, and the proposed analysis reveals that when assertive existential quantification performs in situations, assertions are sensitive to (i) the lexical aspect of the verbal predicates and (ii) the specificity of the situation. It leads to the theoretical implications that: (i) endpoints/boundaries of situations should be linguistically differentiated between actualized and potential ones, following Dahl's (1981) proposal; and (ii) stage-level stative patterns are more like events than states in its endpoint properties.

This paper presents a semantic analysis of the perfect aspect in Niuean (Polynesian). We argue based on novel data that the Niuean perfect shares many properties with the English perfect, but also several striking differences. The Niuean perfect disallows universal perfect readings, can induce inchoative readings, and with some event-types allows present in-progress readings. We analyze the Niuean perfect as introducing inchoative semantics: it adds an initial change-of-state to the predicate to which it attaches, and places that change-of-state inside the Perfect Time Span. Our findings cast doubt on a model of grammar whereby lexical aspectual operations such as inchoativization are strictly separated from viewpoint aspectual operations such as the perfect. We argue for a flexible approach whereby different components of meaning (inchoativity, a Perfect Time Span restriction, current relevance effects) can be introduced at different levels and combined in different ways by languages to produce surface cross-linguistic variation.

The imperfective paradox (IP) refers to the fact that the imperfective-progressive yields completion entailment with atelic predicates (e.g., Livia was pushing the chair → Livia pushed the chair = true) but not with telic predicates (Livia was peeling the tangerine → Livia peeled the tangerine = not necessarily true). The paper questions whether L2 learners too – like adult native speakers – are sensitive to the IP. Ninety-nine adult L2 Italian learners with different L1s and proficiency levels underwent a novel version of the completion judgment task based on event sub-intervals. In this task, learners did not assess if imperfective and perfective events in a video clip was completed but when. Analysis of reaction times and of interruptive-clicks showed that beginner and intermediate L2 learners – unlike native speakers – did not differentiate the patterns of completion between perfective and imperfective predicates and between telic and atelic predicates. It is possible that – at initial stages of acquisition – such aspectual oppositions are still underspecified. These results challenge the predictions of the Lexical Aspect Hypothesis.

This article presents the verbal suffix -ite in Totela (Bantu, Namibia and Zambia), which has variable temporal interpretations based both on lexical aspect (situation type) and on discourse context. The same -ite-marked predicate may be interpreted as referencing a situation that is past (resultative-like readings) or present (progressive-like readings) with respect to utterance time. The suffix is analyzed as having the aspectual function of a stativizer, asserting a relevant property of the utterance's subject. Temporal interpretations with respect to utterance time (or other perspective time) are derived from principles of relevance: the state described by an -ite-marked predicate is interpreted so that it answer the current question under discussion in discourse. The -ite suffix, most likely related to a historical resultative, still specifies a result state, but the temporal specifications of that result state are weakened and must be inferred through context. Cross-linguistic comparison of -ite with other markers that have both perfect/resultative and progressive readings suggests that the pragmatic notion of relevance ⿿ and not only commonalities in temporal semantics (e.g. focus on post-state; stativizing functions) ⿿ may be a key factor in the perfect/resultative/progressive connection.

This paper investigates the degree to which, in a Moraic Phonology framework, coda weight can be seen as responsible for nuclear length in several varieties of French. First, I outline the distribution of vowel length as detailed in conservative variants of French (Passy, 1913; Grammont, 1928). The proposed analysis discusses several length-related processes that are claimed to result from a very simple operation: the removal of an association line that links the moraic tier to certain terminal segments in the melody. Extraprosodicity at the word-level is shown to play a crucial role in determining which elements qualify as mora-bearing units. The second part carries the same analysis into the realm of a variety of regional French, that of Lower Normandy (LNRF), where length contrasts regulate a large part of the phonology. Two features stand out as separating LNRF from Conservative French: the lexical aspect of length — considerably more enhanced in LNRF — and the relationship between nucleus length and the weight of a subset of latent coda consonants — sporadic in Conservative French, systematic in LNRF.