DIALOGUE AND DIALOGIC STRUCTURE: ENHANCING INDIVIDUAL REASONING THROUGH ARGUMENTATION

Mercier and Sperber (Mercier & Sperber, 2011) proposed a new theory of reasoning in 2011: the Argumentative Theory, which asserts that the function of reasoning is argumentation and that the use of reasoning is best explained as a consequence of an evolution in communication. According to Mercier and Sperber, because reasoning evolved to produce and evaluate arguments, reasoning should perform well in this context; second, if the goal of reasoning is to communicate, the ideal context for reasoning to perform well should be the argumentative context, which can be provided by group discussions. The two authors acknowledge that groups are not exempt from social and cognitive biases, but if certain characteristics are maintained (e.g., divergence of opinions, greater focus on the evaluation of arguments rather than on production, and accuracy orientation) then group discussion can lead to superior performance. Several studies corroborated this conclusion (Claidière et al., 2017; Mercier, 2016; Mercier et al., 2016; Mercier & Sperber, 2011; Trouche et al., 2014). However, some questions remain unanswered, such as the replicability of effects of group discussion using tasks without a demonstrably correct solution, or the identification of the underlying cognitive mechanisms of group debate effect. The present thesis focused on these issues, and four studies were conducted to provide some answers. The first two studies investigated group argumentation and its effect on individual reasoning in misinformation and falsification tasks. The findings of both experiments verified the group's positive effect on individual reasoning. In study 1, individuals improved their ability to recognize fake news only when the news was discussed in the group, not when they had to write an individual argument. Study 2 demonstrated that group discussion enhances individual falsification skill even when no one in the group proposes the correct answer, and that accuracy can only be obtained through debate. Studies 3 and 4 focused on the dialogic structure, examining the effects of a mere exposure to a written dialogue. The findings demonstrated that exposure to a dialogic structure (i.e., without actual participation in a discussion) is sufficient to enhance individual reasoning, resulting in significantly higher performance than a control condition with the same arguments but in list format. The dialogic structure proved effective in both news evaluation and falsification tasks.