tv.um.es

The main aim of our research is to focus on the argumentative patterns specific to different
semantic types of modality. As modal forms do not directly contribute to the semantic content
of an utterance, their analysis is better documented if we consider the rhetorical level of
discourse. Thus, the modal forms we examine are analyzed in relation to the connectives that
give rise to particular argumentative patterns, reflecting the rhetorical level of discourse. Our
theoretical background consists of the wealthy literature highlighting the polysemy and
polyfunctionality of modal forms. The study of modality displays a rich amount of theoretical
backgrounds (such as cognitive pragmatics – Saussure & Barbet 2012; evidentiality framework
– Dendale & Tasmovsky 2001, Desclés & Guentcheva 2001; formal semantics – Kaufmann &
Kaufmann 2015, Kratzer 1991) to account for the diversity of semantic flavors that a modal
form can take, depending on the context in which it occurs. Our purpose is to propose a new
framework according to which the argumentative patterns in which a modal form occurs (i)
predetermine its semantic type and (ii) shape the different flavors that the form endorses. Using
a corpus-linguistic approach with statistical tools and methods, our results show that a modal
expressing necessity such as devoir and a modal expressing possibility such as pouvoir have
symmetric behaviors with connectives expressing purpose – devoir is specific to the left of the
connective and anti-specific to the right, whereas pouvoir is specific to the right of the
connective but not to the left. The argumentative pattern expressing concession with a
connective as mais shows similar tendencies: devoir is anti-specific when it occurs before mais
and highly specific when it occurs after, whereas pouvoir does not show such a discrepancy
between the left or right position, being specific in both places. As for the flavors that a modal
verb can endorse, we also observe differences concerning the right or the left position of the
modal in relation to the connective. Pouvoir can support a capacity/permission value (root
modality) or an epistemic value. In the first case, it occurs slightly more frequently to the right
of mais and in the second case to its left. Such a tendency becomes clearer with the epistemic
adverb peut-être.
Our methodology is partially corpus-driven. For each modal verb, we identify the specific
connectives in corpora representing two genres of informational discourse – press and
encyclopedia – within asymmetric spans (5 to the right and 10 to the left) since the modal verb
is more likely to occur just after the conjunction due to the French syntax. We use statistical
measures giving a score according to which the cooccurrence of the connective and the modal
verb cannot be due to chance. This first step allowed identifying the collocational profile that
particularizes devoir and pouvoir, cf. (i). Then we went more in-depth into the analysis of this
collocational profile, taking into account the semantic values of pouvoir. We calculated the
percentage of epistemic pouvoir and the specificity score of peut-être before and after mais, cf.
(ii). We will develop both hypotheses considering other argumentative patterns and modal
forms and coding the semantic flavors for each modal form in our different corpora to see if
they bring out similar results.