Gascon, a dialect of Occitan spoken mainly in Southwestern France, is the only Romance language to possess an elaborate paradigm of preverbal particles traditionally called "enunciatives" (énonciatifs) in linguistic literature. The most frequent element of this paradigm is preverbal que, which in contemporary Gascon surfaces in most main clauses. Preverbal e, the second element of the paradigm, is much rarer in spoken language due to sentence-phonetic restrictions. Furthermore, one may find a significant number of examples where there is no enunciative particle at all, although sentence structure and sentence phonetics would allow for the insertion of such a particle.

Gascon enunciatives share some properties with 'classic' discourse particles, for they do not have any semantic weight derived from their lexical meaning, nor do they contribute to the propositional meaning of the sentence or clause the verb of which they precede. Like discourse particles, enunciative particles are not totally obligatory - as has just been pointed out -, although preverbal que is much more grammaticalized than discourse particles in Romance languages normally are. Enunciatives differ from 'classic' discourse particles insofar as they may not be shifted within the utterance but occupy a fixed position and may be separated from the finite verb only by clitic pronouns. Enunciatives - again unlike discourse particles in many languages - may not cluster in their preverbal slot but are mutually exclusive. On the other hand, it is interesting to note that these enunciative particles, at least in some subdialects of Gascon, may not co-occur with certain typical discourse particles such as desemantisized intensifying adverbs.

The genesis of the enunciatives' paradigm and the etymological origin of its elements que and e are probably impossible to reconstruct on empirical grounds due to a lack of historical textual evidence. It seems reasonable, however, to consider preverbal que as the starting point from which the paradigm evolved. Que, the pan-Romance complementizer, must obviously have some constructional origin, possibly a cleft construction used not for focus purposes but for clause linkage.

Despite this diachronic uncertainty and notwithstanding the restrictive structural features by which enunciatives differ from prototypical discourse particles, linguists who dealt with Gascon morpho-syntax from a functional point of view think that these preverbal markers in present-day usage have two functions normally associated with discourse particles: to structure the discourse according to theme-rheme-oriented criteria and to modalize the utterance. In this talk, I would like to focus the modalizing function of enunciative particles. Pragmatic interpretation of enunciative distribution in spoken and literary corpora suggests that these Gascon particles function as attitudinal markers, expressing the assertive value that the speaker wants to attribute to the utterance. In the light of such a pragmatic interpretation, preverbal que signals unrestricted assumption of communicative responsibility by the speaker, whereas preverbal e expresses reduced assertion. No enunciative particle is introduced where assertive values are not at stake. Therefore, with its preverbal particles Gascon seems to dispose of a supplementary modalizing device, as this Occitan dialect - like all other Romance languages - may also express different assertive modalities by verbal inflectional devices (use of indicative vs subjunctive mood). From this point of view, it is tempting to ask if the enunciatives may also function as markers of evidentiality, a pragmatic category by which the speaker expresses - sometimes by morphological or morpho-syntactical means - where his knowledge of the state of affairs he talks about comes from. This pragmatic category is therefore closely related to the expression of assertive values. Again, Romance languages, which in general have only poorly morphologized devices to signal evidentiality, do so through verbal mood, but in Gascon there are enunciative uses which may be seen as instances of evidentiality marking.

A last remark of the talk will be devoted to the question on how on-going grammaticalization reduces the aforementioned pragmatic value of the enunciative particles.