Blanchette, I. and Richards, Anne (2004) Reasoning about emotional and neutral materials: is logic affected by emotion? Psychological Science 15 (11), pp. 745-752. ISSN 0956-7976.
Abstract
In two experiments, we investigated whether people reason differently when they reason about emotional and neutral contents. We gave participants a conditional reasoning task (“If p, then q”) and varied the emotionality of the items used as p and q. Participants were asked to draw inferences based on these statements. In Experiment 1, we compared statements including preexisting emotional and neutral words. In Experiment 2, we experimentally manipulated the emotionality of initially neutral words using classical conditioning. In both experiments, emotionality affected participants' responses. They were more likely to draw invalid inferences in response to emotional compared with neutral statements. When people reason about important issues, emotions are likely to be involved. Be it personal relationships, important social issues, or material possessions that cost a lot of money, people frequently reason about things that have strong emotional value. Are there differences between reasoning about emotionally involving situations and neutral situations? Empirical evidence concerning the way in which emotions affect reasoning in nonclinical populations is scarce. We investigated this question in two experiments. There is a commonsensical notion that emotions have the potential to impair logical reasoning, that a cool head is more rational than a hot head. This idea may find its roots in several important philosophical traditions. Dating back to the Stoics, various philosophical theories have portrayed emotion and cognition as working in opposition, with passions having a disruptive influence on reason and rationality (De Sousa, 1987; French & Wettstein, 1998; Lyons, 1980). Some early psychological theories adopted similar assumptions (Dewey, 1895; Hebb, 1949). Current perspectives in emotion research offer a radically different view. Contemporary approaches to the study of affect in psychology and the neurosciences have consistently emphasized the adaptive value of emotion (Ekman & Davidson, 1994; Frijda, 1994; Keltner & Gross, 1999; Levenson, 1994). Widespread in current research is a functionalist approach, based on the notion that emotions serve important roles and that they provide benefits to individuals and groups who experience them. Interactions between cognition and emotion have frequently been studied within this framework. One type of empirical evidence supporting the functionalist view concerns the relation between decision making and the experience of affective states. Neuropsychological studies show that people who are unable to experience emotions, but have preserved cognitive skills, may be seriously impaired on certain decision-making tasks (Damasio, 1995; Dimitrov, Phipps, Zahn, & Grafman, 1999). Thus, emotions may promote sound thinking rather than hinder it. The question of whether emotion affects reasoning actually encompasses two related but separate questions. One is whether the affective state currently experienced influences reasoning, independently of what one is reasoning about. The other is whether people reason differently about emotional and neutral content, independently of their current affective state. There is some empirical evidence concerning the former question, which we describe next, but little evidence bearing on the latter question, which is the focus of the two experiments reported in this article. Oaksford and his colleagues (Oaksford, Morris, Grainger, & Williams, 1996) have provided evidence that affective state, or mood, affects reasoning behavior. They manipulated participants' mood and looked at the effect on the Wason selection task.1 Participants in both positive and negative moods were less likely to provide normatively correct answers, compared with participants in a neutral control group. Although this study showed similar effects of positive and negative moods on reasoning behavior, some studies have provided evidence that the specific mood experienced may lead to a specific style of reasoning. For instance, research on social cognition shows that positive moods generally promote the use of top-down strategies, whereas negative moods increase more systematic processing (e.g., Fiedler, 2000). This finding is consistent with a functional view of emotion, according to which specific emotions promote particular ways of thinking, or increase the likelihood of drawing certain inferences. Thus, there is evidence that affective states influence reasoning; however, little research has examined whether the emotionality of the content has an effect. The present experiments explored whether people reason differently about emotional and neutral content, independently of their current affective state. Research on reasoning has shown that content has an important effect on the type of inferences that will be drawn (for reviews, see Evans, Over, & Manktelow, 1993; Manktelow, 1999). A number of content variables have been systematically investigated, including familiarity with the rule2 (Cheng & Holyoak, 1985), agreement with the conclusions (Evans, Barston, & Pollard, 1983; Markovits & Nantel, 1989), and perceived necessity and sufficiency (Cummins-Dellarosa, Lubart, Alksnis, & Rist, 1991; Markovits, 1984, 1986; Thompson, 1995, 2000). Thus, there is evidence that content affects reasoning; however, most of the content variables investigated have been cognitive in nature. There is research showing that people are more likely to endorse attitudinally congenial conclusions as logically valid than they are to endorse conclusions with which they disagree (Henle & Michael, 1956; Morgan & Morton, 1944). Attitudes typically involve affective components. However, attitudes also involve cognitive components, and these studies do not allow us to differentiate between the affective and cognitive components of attitude congruence. The present experiments focused specifically on the emotional value of the reasoning contents so we could determine whether emotional content affects reasoning. To investigate this question, we used a conditional reasoning task. Conditional reasoning involves statements of the type “If p, then q.” Although it is formalized in propositional logic, this type of reasoning is prevalent in everyday situations. From a conditional statement such as “If a person rides a bicycle, then she burns calories,” various inferences can be drawn; some of these are logically correct, and some are not. Four possible inferences are as follows: Modus ponens (MP) takes the form “p, therefore q”: “Chris is riding a bicycle; therefore, she is burning calories.” Modus tollens (MT) takes the form “not q, therefore not p”: “Chris is not burning calories; therefore, she is not riding a bicycle.” Denying the antecedent (DA) takes the form “not p, therefore not q”: “Chris is not riding a bicycle; therefore, she is not burning calories.” Affirming the consequent (AC) takes the form “q, therefore p”: “Chris is burning calories; therefore, she is riding a bicycle.” Logically speaking, the first two inferences (MP and MT) are correct, but the latter two (DA and AC) are not. It is logically possible that Chris is not riding a bicycle but is burning calories nevertheless (thus, to infer “not p, therefore not q” is incorrect). Similarly, it is possible that Chris is burning calories but not riding a bicycle (thus, “q, therefore p” is also incorrect). Because conditional reasoning is both commonly used and the subject of rigorous prescriptions based on normative logic, it provides a formal and relevant framework for studying the effect of emotional content on logical reasoning.
Metadata
| Item Type: | Article |
|---|---|
| School: | Birkbeck Faculties and Schools > Faculty of Science > School of Psychological Sciences |
| Depositing User: | Sarah Hall |
| Date Deposited: | 21 Jan 2020 15:50 |
| Last Modified: | 02 Aug 2023 17:57 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/30650 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
