Perfect Semantics. How universal are Ibero-American Present Perfects ?

Abstract

The semantics of perfects presents a number of interrelated puzzles which have received a lot of attention in recent years. At the heart of these puzzles lies the fact that Present Perfects usually exhibit a restricted distribution when compared to other perfect forms. This distribution can be described as an array of different 'readings' and it is subject to a considerable degree of cross-linguistic and dialectal variation. Portuguese and some American Spanish varieties make a very restricted use of the Present Perfect, and they are widely held to instantiate a stage of development in which Present Perfects only exhibit readings (Harris 1982, Squartini & Bertinetto 2000). On the basis of original descriptive work on a Brazilian Portuguese variety and of a comparison with data on two American Spanish varieties, Mexico and Rio de la Plata Spanish, this paper shows that the Portuguese and the (restricted) Spanish Present Perfects differ in some crucial respects, and calls into question the assumption that readings constitute a stage in the development of perfects. Rather, what the varieties we study have in common is the fact that the Present Perfect has made little or no headway in the competition against the older, simple perfect form ('simple past'). In Portuguese, this has resulted in a special Present Perfect with a clearly profile and a number of pluractional characteristics. In the restrictive American Spanish varieties, by contrast, the Present Perfect exhibits both and existential/resultative readings, but it widely lacks obligatory contexts of occurrence and is not felicitous for reference to singular past events or their results. This paper is also an attempt to bring nearer to each other two trends of research that tend to live in mutual ignorance: formal approaches to semantics, on the one hand, and descriptive research on grammaticalization, on the other. An enhanced awareness of the extent of cross-linguistic variation and of the diachronic processes involved in this variation could undoubtedly help formal linguists determine the point at which it is advisable to give up the search for cross-linguistically uniform meanings. Conversely, the standards of expliciteness associated with formal linguistics could certainly contribute to a refinement of the generalizations proposed by grammaticalization theory. The paper is structured as follows. In section 2, I will recall the different readings associated with well-behaved Perfects, illustrating them with (Standard-) European Spanish, and I will briefly review the rival formal accounts for Perfects. In so doing, I will try to bring out as precisely as possible what a universal reading is and under which conditions it surfaces. In section 3, I will take up the issue of variation and of non-well behaved Perfects. I will first explore those Present Perfects that are ambiguous between a perfect and a simple past configuration, and I will show that contemporary European Spanish usage instantiates this type. Section 4 is devoted to Portuguese and to the IberoAmerican varieties whose Perfects are not well-behaved on account of their inability to be freely inserted in existential or resultative contexts. Section 5 concludes on a speculative note as to the differences between Portuguese and Spanish.