The aim of the programme is to analyze the meaning of the perfect. This sounds easy, but it is not. Verb forms like English ‘have left’ and their counterparts in other languages combine reference to the past and the present, in ways that have escaped linguistic theory. To get out of the doldrums, we need data that enable cross-linguistic comparison without pre-conceptualized meanings. We obtain those data through a new methodology for multilingual corpus research that we dub ‘Translation Mining’. As the perfect interacts with the grammatical categories past and present, its analysis requires a competition-based approach. Our corpus data show the alternations, so we improve on current analyses by developing a micro-typology that is grounded in translation equivalences between English, Dutch, German, French and Spanish. Perfects vary not only in sentence-level meanings, but also in narrative and non-narrative discourse use. As a theory of temporal structure in non-narrative discourse is lacking, we build a temporal extension of commitment-based dynamic theory to analyze the corpus data of perfects in context. The predictions of a corpus-based analysis need to be checked on an additional corpus and squared with native speakers, so we carry out a corpus experiment and native speaker surveys. The outcome of the programme will be (1) a semantic map of the perfect, (2) a dynamic account of its non-narrative nature, and (3) a unified truth-conditional and dynamic semantics of its sentence-level and discourse-level meanings. The results will be made available for translators, computational linguists and typologists. 

You can meet our team and get acquainted with our research through the following (3 minute) video: https://vimeo.com/222494190. 

Further information is available through the project's website, see: http://time-in-translation.hum.uu.nl

• prof. dr. Henriette de Swart
• Martijn van der Klis MSc
• dr. Jos Tellings
• dr. Martin Fuchs

This project brings together formal semantics research on article systems and possession with L2 research on the acquisition of referentiality. For more information, including information about the project's workshops, please visit my personal webpage.

Articles and bare nominals are often in complementary distribution. English tolerates bare plurals and bare mass nouns (buy apples, drink milk), but not generally bare count singulars (*buy book), except in special configurations like at school, in bed, without anchor), in which the bare nominal is not so much to refer to a concrete individual, but rather to describe properties, types, conventional situations, etc. The use of articles in such weakly referential constructions is often variable within one language (be chair of the committee vs. be a professor), as well as across languages (be a professor - être professeur). The project works out three complementary questions in relation to the distribution and meaning of weakly referential expressions:

(i) weakly referential nominals: lexicon and constructions. Research on classes of nouns and weakly referential configurations.

(ii) Weakly referential nominals: meaning and comprehension. Online/offline experimental investigation on the way in which weakly referential nominals are understood.

(iii) Cross-linguistic distribution of weakly referential nominals. Corpus research on different languages and parallel corpora (English, Romance, Germanic) provides the data for projects (i) and (ii), as well as the generalizations that support information extraction, computer aided translation and natural language processing.

• prof. dr. Martin Everaert
• prof. dr. Henriette de Swart
• dr. Joost Zwarts