Aspect and Temporal structure
The domain of aspect and temporal structure in learner language is investigated at early stages of the development of lexical and morpho-syntactic means and at very late stages of acquisition, i.e. beyond the acquisition of form as such. In both cases, work is carried out in the theoretical framework developed by W. Klein in Time in Language (1994).
(1) Investigation of the early stages of development deals with four major research questions, which are currently underway.
A first question has to do with the linguistic motivation underlying the development of morpho syntactic forms. Prior findings based on longitudinal studies (ESF, Pavia) suggested that development in the use of tense and aspect marking takes place when the linguistic means to cope with specific discourse contexts are not available. In this line, individual studies focus on contexts which attract tense and aspect distinctions as for example foregrounding and backgrounding information in narratives (U. Paris 3), in reported speech (U. Nantes) or in impersonal and passive constructions as means to mark different perspectives on an event (U. Pavia). Clear correlations of impersonal and passives with aspectual and modal properties have been noted: si-constructions have predominantly imperfective aspect when used with present and imperfect, while periphrastic passives often have perfective aspectual value.
A second question deals with the relative impact of the inherent temporal properties of the verb on the use of tense and aspect. The U. Pavia has undertaken an investigation on Actionality and Aspect in Second Language Acquisition based on corpora of Italian as a second language, including both spontaneous and tutored learners. Results suggest a partial revision of the way the prototype notion has been applied to Second Language Acquistion (by Andersen and Shirai, in particular). Learners do not seem to be guided by any 'Innate predisposition" to recognize the inherent properties of predicates ; they rather follow an inductive procedure and assemble appropriate features, such as adverbials, direct objects, at verb phrase level. This hypothesis would lead to a re-evaluation of syntax in the acquisition of tense and aspect systems. First results were presented at international meetings (Bergamo meeting among others) and discussed in the framework of an Italian research project on first and second language acquisition financed by the Ministry for Research and Education.
Collaborators directly involved in the project: Rosi, Fabiana fabiana.rosi/AT/gmail.com ; Rastelli, Stefano srastelli/AT/tiscali.it ; website for Giacalone Anna,with downloadable papers.
(2) Cross-linguistic studies carried out in close cooperation in the University of Heidelberg, Paris VIII, and Radboud University Nijmegen focus on very advanced L2s and investigate the extent to which language-specific preferences in information structure and temporal reference are driven by grammaticised means.
Linking events in sequence
Given the nature of the task studied (narrative/silent film), the underlying principles determining information structure have a multidimensional complexity since information organisation spans conceptual domains such as time, space, entities, events. This means that adult learners have to acquire the macrostructural principles in the target language that lead to coherence building on a large scale.
Focusing on the hierarchical status of tense and aspect in narrative tasks, and the means implemented in creating coherence and setting up temporal frames of reference, comparisons of Dutch, English, French, and German present evidence for a hierarchy of factors and associated constraints that are both grammatically based and perspective driven (Carroll, Rossdeutscher, Lambert, & von Stutterheim 2007). They constitute hierarchically ordered principles that guide the decisions made in information selection and information organisation (information selection ‘deciding what to say’, referential framing (anchoring predicate argument structures with respect to worlds and spaces), selection of a temporal frame/temporal relations, topic management). Given its complexity, the underlying linguistic knowledge provides a window on factors relating to ultimate attainment in adult second language acquisition. The crosslinguistic comparisons of advanced L2s cover temporal frames of reference in narratives (von Stutterheim & Lambert 2005; Carroll & Lambert 2006; Carroll & von Stutterheim 2003; van Ierland & Starren 2007), information structure/information selection (Carroll & Lambert 2003), topic management (von Stutterheim & Carroll 2005; Carroll 2007).
The findings show that on the whole advanced learners retain hierarchical orders associated with grammaticised features of the L1. However, while very advanced learners (near native) are derailed, at the final stages, by what can be called structural similarities of a subtle nature (the syntactic subject in preverbal position is relevant in information structure in both English and German, for example, but has a different hierarchical status in English compared to German), less advanced learners do not recognise the full implications for information structure, in macrostructural terms, of a range of grammaticised means of the target language. We assume that at very advanced stages ultimate attainment is not hindered by grammatical knowledge as such, but by its implications for information structure – as expressed in macrostructural principles and related hierarchies. The question is how do learners derive the appropriate macrostructural knowledge from the input and learn to put it into practice.
Decontextualised events: Specificity and temporal categories
Based again on hypotheses related to grammaticised categories, a series of crosslinguistic studies were carried out which involve the verbalisation of individual events in decontextualised situations (Germanic: Dutch, English, German; Romance: French, Italian, Spanish; Slavic: Czech, Russian, Polish; Semitic: Modern Standard Arabic, Algerian Arabic). Using the same set of tasks (40 scenes/everyday situations), the studies focus on the question of specificity and how speakers proceed when grounding an event as an individual case. The relevant linguistic means used in anchoring information in the domain of discourse (tell what is happening) relate to clause structure, tense and aspect, verb selection, use of complements and adjuncts, as well as definite and indefinite reference. Analyses of the L1s provide insights into the status of finiteness (assertion time), aspectual distinctions (progressive, perfectivity, boundedness) and the means used in reference introduction (existentials vs. main clauses) in determining verb selection and the use of adjuncts when anchoring an event as a specific case. Ultimate attainment in L2 acquisition is again investigated in terms of underlying decision hierarchies and associated grammaticised means.
In an additional study on progressive aspect, languages were compared in which use of the progressive in telling what is happening is not obligatory. In Dutch, French, Italian, and MSA, for example, the progressive (aan het zijn; stare + gerund; en train de) constitutes an option in telling ‘what is happening’ and anchoring an event as an individual case. In testing the factors which lead to use of the progressive, the study reveals that use of this aspectual distinction is dependent on a similar set of cognitive/semantic features across the different languages.
As a window on planning processes and language specificity, a series of experimental tasks (eye tracking, speech onset) were designed to test the extent to which grammatically driven preferences in information selection and information organization are reflected in factors such as 'focus of attention' while processing information for expression (intake). In addition, a time pressure experiment was carried out (shorter clips, shorter time span for verbalization) in order to test the status of grammaticized means in event construal.