Biased communication: The Cognitive Pragmatics of Fallacies
This interinstitutional project aims at investigating persuasive, and in particular manipulative, exploitation of otherwise efficient cognitive pragmatic processes of understanding. This research is anchored on the theoretical vantage point of pragmatics, which predominantly depend on models of understanding relying on frugal (rapid) but non-prudent (risky) processes of contextualisation and inference, which are compatible with spontaneous and automatic heuristics documented elsewhere in cognitive psychology; we draw upon recent developments in this framework aiming at bridging the explanatory gap between understanding and consenting through the notions of epistemic vigilance (Sperber & al. 2010), context-selection constraint (Maillat & Oswald 2009, 2011, Maillat 2006, forthcoming, and previous work by Saussure (2005) and others. The outcomes of the project, besides a better understanding of human sensitivity to fallacious arguments, lie on a more practical level in establishing better control procedures, that is, a critical mind, in the greater public. In this project, we approach the interpretative processes triggered by fallacious arguments and their persuasive and manipulative efficiency, from the theoretical vantage point of pragmatics which considers human communication as a cognitively driven activity which tries to maximise the output of the interpretative process and simultaneously minimise the amount of resources summoned during this same process. In brief, the current project looks at how, through fallacies, persuasive and manipulative discourse exploits cognitive biases which hinder this interpretative process and yield sub-optimal, or even irrational, outcomes. In doing so the main goal of this project is to bring the centuries old discussion of argumentative moves and fallacious moves in particular (van Eemeren & Grootendorst 2004, Hamblin 1970, Walton 1995, 1996, 2000, 2003, Woods & Walton 1982) into the domain of linguistic – more precisely pragmatic – theory in order to move away from a mostly descriptive approach to such discursive phenomena to an explanatory approach that will use the pragmatic theoretical framework in order to make predictions regarding the comprehension processes at work when an addressee interprets well-known fallacies, for instance the ad verecundiam, ad populum, ad hominem, ad baculum, strawman, ad consequentiam, etc. In this respect, this project fills an important part of the gap noted by Cummings (2004) who emphasizes the lack of and need for a theoretically grounded pragmatic account of argumentative moves. As explained above, this project tackles the depth and scope of pragmatic enrichment processes through contextual selection and modulation (disambiguation, reference assignment, semantic saturation of elliptic forms) in persuasive circumstances (Carston 2002, Sperber & Wilson 1995, Recanati 2007, 2010 and others). The purpose of the first level of investigation is therefore to further develop a model that can capture phenomena that pertain to biased communication – namely, fallacious arguments – and which is inscribed within the larger framework of pragmatics, thereby pursuing and extending the initial theoretical steps taken by Maillat & Oswald (2009, 2011) and providing an explanatory account which is cognitively grounded. The data used to test the various theoretical hypotheses is drawn from an evaluation of the comprehension processes triggered in an addressee by a subset of fallacious arguments. From a methodological point of view, the testing of our hypothesis is done in two distinct and complementary experimental strands (see Noveck & Sperber 2004, Pohl 2004, Sauerland & Yatsushiro 2009). Thus, the first line of investigation focuses on fallacies from the perspective of Context Selection Constraint (CSC; see Maillat 2006, forthcoming and Maillat & Oswald 2009, 2011, forthcoming) a pragmatic account of biased communication that was specifically developed to capture fallacious arguments. Specifically, the project experimentally test the biases theoretically predicted to be prompted by four fallacies: ad populum, ad verecundiam and ad baculum in interpretative processes. In doing so these fallacies are systematically and respectively related to the relevant cognitive counterparts the mere exposure and validity effects, epistemic vigilance, and somatic markers (see below for a discussion). Thus we use the insights of cognitive psychology in order to test empirically the validity of our predictions. Indeed, one of the original contributions of this project rests in its interdisciplinary effort to bring together the findings of scholars who investigated fallacies with those put forward by people who – following the ground breaking work of Tversky & Kahneman (1974, 1981), explained some of the most puzzling aspects of human understanding, judgment, and decision making (see Pohl (2004), or Gigerenzer (2008) for recent surveys). Interestingly, the relevance-theoretic framework in which the theory is couched stimulates and assumes the parallel between general cognitive processes and pragmatic processes as the latter are taken to exploit generally valid cognitive principles (see Sperber and Wilson 1995), thereby supporting the combined approach discussed above. The second line of investigation concerns the so-called strawman fallacy where the speaker gets attributed a commitment to a content she does not (intend to) convey. The research question concerns the efficiency of this fallacy in persuasion, an issue expanding far beyond the question of persuasion, having to do with the overall mechanism of pragmatic inference and of retractability (itself a criterion for implicitness in classical Gricean-style pragmatics). In argumentation theories, expressed contents are reputed public and scrutinisable, objective facts, a standpoint we regard as a rough oversimplification relying on an ill-informed theory of language understanding where the role of inference is null or light. Furthermore, the classical assumptions on this matter by Argumentation theories cannot provide any explanation for its common success otherwise than just through the notion of burden of proof switching. On the contrary, the tools designed within cognitive pragmatics and general principles of understanding in context, as they provide explanations for other types of miscommunication, such as misunderstanding or quiproquo, suggest a line of explanation whereby the strawman relies on the higher relevance of the attributed content with regard to the actually intended one. Empirical tests are proposed to evaluate the model.