22 November 1996

Adversativity, Argumentation, and Refutation

Scott Schwenter

Stanford University

In this paper, I explore different types of refutation, concentrating on the contrast between two adversative connectives in Spanish, pero `but' and si as these are used in dialogues. The latter form, the canonical conditional sentence marker, is usually translated by `if', but also has adversative uses translatable by `but' in declaratives. The analysis builds on techniques from Argumentation Theory (Ducrot 1996), and focuses on how linguistic elements constrain expression of points of view in discourse. Both pero and si typically introduce a proposition Q anti-oriented to some other preceding proposition P in the context, but differ in how they orient speaker point of view towards P and/or its potential implicatures. The two forms in this usage are exemplified in (1):

  (1a) A: Juan es un genio. (Vamos a invitarlo.)
         `Juan is a genius. (Let's invite him.)'
       B: (Sí) Pero es un idiota (al mismo tiempo). (No vamos a invitarlo.)
         `(Yes) But he's an idiot (at the same time). (Let's not invite him.)'
  (1b) A: Juan es un genio. (Vamos a invitarlo.)
         `Juan is a genius. (Let's invite him.)'
       B: (#Sí) Si es un idiota (#al mismo tiempo). (No vamos a invitarlo.)
         `(#Yes) SI he's an idiot (#at the same time). (Let's not invite him.)'

The difference between these two connectives can be established on the basis of what it is they object to within the prior context: pero reflects a viewpoint that accepts an argumentative sequence P-->R that precedes it (where R=some implicit or explicit conclusion, like vamos a invitarlo in (1a,b)), but then introduces a "stronger" sequence anti-oriented to the preceding one. On the other hand, si also introduces a "stronger" anti-oriented sequence, but cannot be used to accept the prior argumentative sequence P-->R in its totality, as the #-incompatibilities show. It is however possible to use si where one accepts the explicit propositional content of the prior utterance, but not the inferential conclusion R that follows from this content.

A key finding of this study is that refutation is not limited to propositions, but also to inferences from them, or to combinations of propositions and inferences used in an argumentative sequence. As the Spanish pero/si contrast illustrates, these different types of refutation may be conventionally indicated by different forms, which differ in how they orient points of view to what precedes. A hypothesis for further research is that differences between other markers of adversativity in Spanish, English, and other languages can be explained in a similar fashion.

References

  • Ducrot, Oswald. 1996. Slovenian Lectures/Conférences Slovènes. Ljubljana: ISH.