Loading...
Publication
Contraexemplos e raciocínio dedutivo
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.69 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Este trabalho foi desenvolvido com o objectivo de contribuir para uma compreensão mais
alargada do modo como os sujeitos utilizam os contraexemplos no raciocínio condicional,
quando são utilizados conteúdos que remetem para situações comuns do quotidiano.
Não existem dúvidas de que as pessoas são capazes de recuperar contraexemplos, vários
estudos atestam esta capacidade (e.g. Couto, Quelhas & Juhos, 2010; De Neys & Everaerts,
2008; De Neys, Schaeken & D’ydewalle, 2002; Neys, Schaeken & d'Ydewalle,
2003b;Verschueren, Schaeken, De Neys & d'Ydewalle, 2004). No entanto, a forma como os
sujeitos recuperam e utilizam contraexemplos, carece ainda de esclarecimentos. Para cumprir
este objectivo geral, construímos dois conjuntos de experiências que incidem em dois factores
importantes. Em primeiro lugar investigámos a recuperação de contraexemplos e aceitação de
inferências, com avisos e conselhos. Na experiência 1, verificámos que os sujeitos são
capazes de gerar contraexemplos para ambos os tipos de frase, mas não o fazem com igual
frequência para avisos e conselhos. Em seguida, nas Experiências 2 e 3 investigámos o
impacto da disponibilidade (Experiência 2) e da recuperação (Experiência 3) de
contraexemplos, nas inferências que os sujeitos fazem, tendo concluído com alguma surpresa
que este impacto é muito pequeno ao contrário do que se sabe acontecer com condicionais
causais (e.g. Byrne, Espino & Santamaria, 1999; Couto, Quelhas & Juhos, 2010; Cummins,
1995; Cummins, Lubart, Alksnis & Rist, 1991). Ainda no primeiro no conjunto de
experiências, avaliámos os padrões de interpretação que os sujeitos fazem com base nestas
condicionais (Experiência 4), tendo concluído que existe uma variabilidade grande das
interpretações que os sujeitos fazem.
No segundo conjunto de experiências, recorremos a frases causais para estudar o impacto da
recuperação de contraexemplos adicionais. A Experiência 5 demonstra que a recuperação de
contraexemplos é um processo pouco fluente e que esta sensação de falta de fluência afecta o
valor que os sujeitos atribuem aos contraexemplos, conduzindo a um padrão de supressão de
inferências contrário à informação que foi recuperada. A Experiência 6 esclarece que o padrão
de supressão reportado na Experiência 5 se deve às dificuldades de recuperação dos
contraexemplos. Quando estes são fornecidos aos sujeitos, em vez de recuperados da
memória, o padrão de supressão corresponde ao conteúdo que foi apresentado, ou seja, mais
contraexemplos conduzem a maior supressão. Na Experiência 7 confrontámos
contraexemplos e a frequência de Excepções, tendo concluído que os sujeitos parecem
preferir a informação probabilística, à informação que decorre dos contraexemplos.
No geral, os nossos resultados mostram três factos importantes. Em primeiro lugar que o
conhecimento que é recuperado durante o raciocínio tem diferentes funções para diferentes
tipos de condicionais, isto é, os contraexemplos recuperados para Advice têm uma função
diferente dos contraexemplos recuperados para condicionais causais. Além disto, mostra
ainda que o processo de recuperação de contraexemplos é pouco fluente e as pessoas parecem
preferir utilizar a informação probabilística. Por fim, esclarece que os dois factores acima
mencionados ajudam a explicar as diferenças que têm sido encontradas na literatura sobre o
peso que cada contraexemplo adicional tem na aceitação de inferências.
We have developed this thesis with the goal of contributing to a larger understanding of the way in which people use counterexamples during conditional reasoning, when they reason about contents that refer to situations that are common on their daily lives. There is no doubt that people are able to retrieve counterexamples, and many studies attest this ability (e.g. Couto, Quelhas & Juhos, 2010; De Neys & Everaerts, 2008; De Neys, Schaeken & D’ydewalle, 2002; Neys, Schaeken & d'Ydewalle, 2003b;Verschueren, Schaeken, De Neys & d'Ydewalle, 2004). However, the way in which people retrieve and use counterexamples is still in need of some enlightenment. In order to attain our goal, we have developed two sets of experiments, which focus on two important factors. We started by investigating the counterexample retrieval and endorsement of inferences for advice conditionals. On the first experiment, we verified that subjects are capable of generating counterexamples to advice conditionals, but they do it differently for tips and warnings. Following this, on Experiments 2 and 3 we evaluated the impact of the availability (Experiment 2) and the retrieval (Experiment 3) of counterexamples, on the inferences that people draw from advice. Surprisingly, we have concluded that this impact is very little, contrary to what has been found for causal conditionals (e.g. Byrne, Espino & Santamaria, 1999; Couto, Quelhas & Juhos, 2010; Cummins, 1995; Cummins, Lubart, Alksnis & Rist, 1991). Finally, on this first set of experiments, we also investigated the interpretations that people make, based on tips and warnings (Experiment 4), and we concluded that there is an enormous variability of interpretations from advice conditionals. On the second set of experiments, we resorted to causal conditionals to study the impact that retrieving additional counterexamples has on conditional inferences. Experiment 5 shows that retrieving counterexamples is not a fluent process, and that this lack of fluency has a deep impact on the value that people attribute to the retrieved counterexamples, thus leading to a pattern of suppression that is contrary to the information that was recovered. Experiment 6 clarifies that the pattern of suppression reported on Experiment 5 is due to difficulties in the retrieval process. When counterexamples are presented instead of retrieved from memory, the suppression of inferences is compatible with the information that was presented to the participants, that is, more contrerexamples equal more suppression of inferences. On Experiment 7, we confronted counterexamples and the frequency of exceptions, and we concluded that subjects show a preference for probabilistic information, rather than counterexample information, when they make conditional inferences. Overall, our results show three important factors. First, the knowledge retrieved during reasoning has different purposes for different sorts of conditionals, i.e., counterexamples recovered for Advice have a different purpose than counterexamples recovered for causal conditionals. Our results also show that counterexample retrieval is not a fluent process and that people prefer to use probabilistic information. The results above mentioned, aid in explaining the differences that have been found about the weight that additional counterexamples have on inference acceptance rates.
We have developed this thesis with the goal of contributing to a larger understanding of the way in which people use counterexamples during conditional reasoning, when they reason about contents that refer to situations that are common on their daily lives. There is no doubt that people are able to retrieve counterexamples, and many studies attest this ability (e.g. Couto, Quelhas & Juhos, 2010; De Neys & Everaerts, 2008; De Neys, Schaeken & D’ydewalle, 2002; Neys, Schaeken & d'Ydewalle, 2003b;Verschueren, Schaeken, De Neys & d'Ydewalle, 2004). However, the way in which people retrieve and use counterexamples is still in need of some enlightenment. In order to attain our goal, we have developed two sets of experiments, which focus on two important factors. We started by investigating the counterexample retrieval and endorsement of inferences for advice conditionals. On the first experiment, we verified that subjects are capable of generating counterexamples to advice conditionals, but they do it differently for tips and warnings. Following this, on Experiments 2 and 3 we evaluated the impact of the availability (Experiment 2) and the retrieval (Experiment 3) of counterexamples, on the inferences that people draw from advice. Surprisingly, we have concluded that this impact is very little, contrary to what has been found for causal conditionals (e.g. Byrne, Espino & Santamaria, 1999; Couto, Quelhas & Juhos, 2010; Cummins, 1995; Cummins, Lubart, Alksnis & Rist, 1991). Finally, on this first set of experiments, we also investigated the interpretations that people make, based on tips and warnings (Experiment 4), and we concluded that there is an enormous variability of interpretations from advice conditionals. On the second set of experiments, we resorted to causal conditionals to study the impact that retrieving additional counterexamples has on conditional inferences. Experiment 5 shows that retrieving counterexamples is not a fluent process, and that this lack of fluency has a deep impact on the value that people attribute to the retrieved counterexamples, thus leading to a pattern of suppression that is contrary to the information that was recovered. Experiment 6 clarifies that the pattern of suppression reported on Experiment 5 is due to difficulties in the retrieval process. When counterexamples are presented instead of retrieved from memory, the suppression of inferences is compatible with the information that was presented to the participants, that is, more contrerexamples equal more suppression of inferences. On Experiment 7, we confronted counterexamples and the frequency of exceptions, and we concluded that subjects show a preference for probabilistic information, rather than counterexample information, when they make conditional inferences. Overall, our results show three important factors. First, the knowledge retrieved during reasoning has different purposes for different sorts of conditionals, i.e., counterexamples recovered for Advice have a different purpose than counterexamples recovered for causal conditionals. Our results also show that counterexample retrieval is not a fluent process and that people prefer to use probabilistic information. The results above mentioned, aid in explaining the differences that have been found about the weight that additional counterexamples have on inference acceptance rates.
Description
Tese de Doutoramento apresentada ao ISPA - Instituto Universitário
Keywords
Contraexemplos Raciocínio dedutivo Raciocínio condicional Counterexamples Deductive reasoning Conditional reasoning