The treatment of tense within LFG has taken place from a primarily morphosyntactic point of view, which suffices with respect to many issues, but is limited. Within the ParGram (Butt, King, Nino and Segond 1999) grammar writing effort, for example, the first treatment of tense strove to encode complex temporal relations such as future-perfect within one feature whose presence was deduced out of periphrastic constructions such as "will have been". That is, the presence of the three elements "will", "have" and "been" in a clause led to the assignment of a feature TENSE FUT-PERF. However, this approach was soon found to have shortcomings. For one, the computational effort to deduce a composite tense is considerable. For another, once one begins to take into account a semantic theory of temporal interpretation such as DRT (Discourse Representation Theory), for example, it is not clear that the deduced compositional tense is indeed useful for the semantic interpretation. The recognition of this problem led to the further recognition of the well-known problem of sequences of tense, and thus the ParGram group backed off the problem entirely and merely confined themselves to registering morphologically clearly recognizable categories as atomic values.

However, this approach then raises the question of what the difference between the tense features recorded at f-structure and the tense features that come as part and parcel of the morphology really is. In particular, this issue arises with respect to the representation of morphology in the context of LFG, i.e, within the context of realizational morphology or the use of m(orphological)-structure as discussed in the Workshop on Morphosyntax in LFG (Berkeley, 2000).

A general intuition would seem to be that the functional information about tense recorded at f-structure should be different from the morphological marking recorded within the morphology (say, at m-structure). Both of these types of information should furthermore be different from a semantic representation of tense. However, as far as I am aware, there has been no systematic examination of these issues.

Further interesting issues are raised by the use of temporal morphology on nominal categories (e.g., Nordlinger and Sadler 2000), the distribution of temporal morphology over several different elements in the clause (e.g., nonconfigurational tense in Wambaya (Nordlinger and Bresnan 1996), or the well-known case of serial verbs which both carry tense morphology). While these phenomena have been treated in terms of c-structure and f-structure analyses, almost no attempt has been made to integrate a semantic analysis, which could possibly contribute to a better and deeper understanding of the issues at hand.

This workshop thus seeks to integrate and discuss the different perspectives on tense: morphological, syntactic (f-structure and c-structure), and semantic.

About the speakers

Mirriam Butt is a Research Scientist at the University of Kontanz. Her research interests include morphosyntax, historical linguistics and computational linguistics. The bulk of her research is on South Asian languages, with a special emphasis on Urdu, though she has also worked on English and German. She has written and edited several books on syntax, semantics and computational linguistics.

Sheila Glasbey is a lecturer at the University of Dundee, Scotland, UK. Her research interests include linguistic semantics (in particular the semantics of tense, aspect, and temporal reference), computational linguistics, and augmentative and alternative communication (AAC). She has published work on the semantics of 'then' and 'when in narrative discourse, on the English progressive construction, and on the semantics of indefinites.

Louisa Sadler is senior lecturer and currently head of department of the Department of Language and Linguistics at the University of  Essex. Her research interests are lexicalist syntactic theories (principally LFG and HPSG), argument structure and  the syntax-lexical semantics interface, Welsh syntax, and Computational Linguistics, including Machine Translation.

Rachel Nordlinger is an ARC Postdoctoral Research Fellow at the University of Melbourne. She has written grammatical descriptions of the Australian Aboriginal languages Bilinarra and Wambaya and has written a book on Constructive Case in Australian languages. Her research interests include descriptive linguistics, Australian Aboriginal languages, morphology, language typology, syntactic theory and Lexical Functional Grammar.