COLING 82, J. Horeck~ (ed.) 
North.Holland Publialu'ng Company 
© Academ~ 1982 
Towards a mechanical analysis of French tense forms in texts. 
Christian Rohrer 
Universit~t Stuttgart 
I) 
In this paper we want to present a system that analyzes tense 
forms and temporal adverbs in texts 2) . For reasons of ex- 
position we restrict the analysis to narrative texts. 
The system compri.:es the following components: 
a) A parser which assigns a tree structure to sentences. 
The fragment which the parser can analyse corresponds roughly 
to the fragment of English described by R. Montague in 'The 
proper treatment of quantification in ordinary English'. Our 
fragment contains however more tense forms, temporal adverbs, 
and temporal conjunctions. In particular we treat the notorious- 
ly difficult pair of tense forms pass~ simple and imparfait, 
frame adverbials like aujourd'hui, hier, demai____~n,ce lour-l~ , u_nn 
~our plus tard, etc.,as well as the .zonjunctions quand,pendant 
~\[u_e,depuis que,a r~~. 
b) The syntactic analysis constitutes the input to the rules that 
derive the corresponding discourse representations. The syntac- 
tic structure detezmines the discourse representation. 
Let D = S i ................. , S n be a discourse of L. 
A D(iscourse) R(epresentation) S(tructure) for D is con- 
structed by reducing the sentences S I ........... , S n in the 
order in which they occur in D. The reduction of a sentence 
S i proceeds through the application of certain DRS-con- 
struction rules which operate on the syntactic analysis of the 
sentence and work so to speak ' from the top down' (thus the 
syntactic analysis imposes a partial order on how the rules 
are to be applied). 
What DRS-construction rule is to be applied depends on the 
particular syntactic formation rule that was used to form the 
syntactic compound that the application of the rule is to re- 
duce. Where one of the immediate components of the compound is 
a singular term, moreover, the choice of construction rule 
will be determined by the type of this term; thus, for in- 
stance there is a different rule for proper names, for inde- 
finite descriptions and for pronouns, xespectively. Similarly 
there are different rules for each of the tenses considered 
here. 
331 
332 CH. ROHRER 
c) In ordinary truth-conditional semantics truth conditions are 
stated for individual sentences: A discourse consisting of a 
set of sentences is true iff it consists of nothing but true 
sentences. The order in which the individual sentences occur 
in the discourse is irrelevant for the truth conditions. In 
this system truth conditions are formulated for a sequence 
of sentences D = $I, ............ , S n in such a way that the 
order in which the sentences occur in the text plays a 
crucial role. The truth of a discourse D in a model M is de- 
fined as the existence of a proper embedding into M of the 
representation of D. 
A major advantage of the present system lies in its ability to 
deal with cases of intersentential and intrasentential anaphora 
in exactly the same fashion. Tense forms and temporal adverbs be- 
have in many respects like anaphoric pronouns. The DRS-con- 
struction rules handle both temporal and pronominal anaphora. The 
system extracts all the temporal information which is contained 
in the input text. At a later stage the information extracted 
from the text will form the basis for a question - answering 
system. An implementation of the system on a VAX 780/11 is in 
progress. 
1) The work reported in this paper was supported by the 
Deutsche Forschungsgemeinschaft grant Ro 245/12. 
2) The system is based on the following papers: 

References

Kamp, H. (1980) 
' A theory of truth and semantic representation', in: 
Groenendijk, J., Janssen, T., Stokhof, M. (eds.), Formal 
methods in the study of language,Amsterdam, p. 277 - 322. 

Kamp, H. (1981), 'Evenements, representations discursives et 
r~f~rence temporelle', Langages , p. 39 - 64. 

Kamp, H., Rohrer C. (1981) 'Tense in texts', Ms. Stuttgart 
