Semântica de frames e tomada de decisão financeira: a representação lexical de esquemas mentais como explicação para os efeitos de framing

Abstract

The present work inquires on the lexical role in framing effects of financial decision-making through the reference and indexation of schematic semantical structures named Experiential Frame (EF). Framing effects is the phenomenon of a decision-maker deciding differently according to the form two identical problems are expressed, a violation of normative rationality. To account for this behavior, we proposed a model based on the hypothesis that identical objects could be represented by multiple frames, having thelexical forms as the role of the prime in a priming-analogous process. This dissertation has the goal of assessing the cognitive aspects of framing effects through the comprehension of lexical and mental processes in decision-making. Specifically, the present work proposes an explanatory model for framing effects based on Cognitive-Functional Linguistics (CFL) a Frame Semantics, comparing the model with recent empirical research, and an experimental design able to explore frame activation to assess the hypothesis and the model. The main research methods employed are theoretical analysis for the model construction and non-systematic literature review for its assessment. This paper presents a conceptual background and the state-of-the-art regarding judgement and decision-making theories, Idealized Cognitive Models (ICM), including schematic structures as image schemas, frames, and scripts, and shared intentionality. The proposed model features the construction of the EF concept as a generic structure that supports both propositional and non-propositional representations, and postulates that (1) communication happens through shared states like emotions and joint attention, (2) lexical units act like an index evoking EFs associated with them, and (3) same objects might be interpreted by different EFs. The first statement finds good support regarding emotional states and mixed results regarding joint attention, while statement 2 finds some support in experiments with Cloze test and research about emotional word, and statement 3 could not be properly assessed. The proposed experiment sets adult L1 Brazilian Portuguese (BP) speakers of any gender as participants, induced frames (or lack thereof) as independent variables, frame activation ratio and latency of successful activation as dependent, and two frames composed each of an action (verb) represented by an animation, an agent (subject, a character), and an object (noun) as dependent variables. The stimuli are categorized in the 3 first sessions, associated into a frame during the 3rd and 4th blocks, and tested in the 5th block. Two applications developed to support the experiment, move.lab and MultiRepTask are introduced, and the goodness of fit of the experiment is acessed. The main remarks of this paper are that the proposed model finds partial support, and the execution of the experiment might help advance the topic.