Inhibitory Control of Thoughts Better Left Unsaid

An experiment explored the hypothesis that inhibitory ability helps people stop themselves from engaging in socially inappropriate behavior. All participants completed a Stroop color-naming task, after which half of the participants were asked to remember an eight-digit number (inducing divided attention). Participants were then offered an unfamiliar and visually unappetizing food product (a chicken foot) under conditions of either low or high social pressure to pretend that it was appealing. Participants who had full attention available and were under pressure to pretend the food was appealing were least likely to emit a negative response, and performance on the Stroop task predicted the degree to which they successfully restrained negative responses. These results suggest that the cognitive ability to inhibit unwanted information facilitates socially appropriate behavior. With all the inappropriate and unfriendly things that people think and say about each other (Rosnow, 2001), how is it that interpersonal interaction is so often positive? What enables translation of socially insensitive or inappropriate cognition into pleasant interaction? One answer to this question focuses on the role of cognitive inhibition in social interaction. Specifically, it may be the case that effortful inhibition of inappropriate but prepotent responses is a critical cognitive component of social skill. The clearest support for this idea can be found in research on individuals who have compromised inhibitory functioning, because they are very young (Kieras, Tobin, Graziano, & Rothbart, 2005), are very old (von Hippel & Dunlop, in press), or have brain damage (Beer, Heerey, Keltner, Scabini, & Knight, 2003). With regard to young children, those who perform better on measures of inhibition are also better able to mask their disappointment when they receive a gift they do not like (Kieras et al., 2005). At the other end of the life span, research suggests that agerelated increases in social inappropriateness are mediated by inhibitory losses (von Hippel & Dunlop, in press). These findings suggest that social control is achieved in part through inhibition, but it is unclear whether inhibition plays a role in social functioning among individuals whose inhibitory faculties are intact. On the one hand, it may be the case that inhibition is largely irrelevant, and that knowledge of the rules of social interaction isall that isnecessary tobesocially skilled.For example, Salovey and Mayer’s (1990) provocative theory of emotional intelligence tacitly equates understanding how to manage emotions with the ability to actually do so in complex social situations. Experimental results support such a possibility, as the emotionmanagement subscale of their emotional intelligence test (Mayer, Salovey, & Caruso, 2002) predicts people’s ability to get along with their friends (Lopes et al., 2004). The roots to this approach can be found in Argyle’s (1969) seminal research on social skills, which suggested that social skills emerge from knowledge of social rules rather than ability to apply them. The success of these major research programs attests to the importance of knowledge of social rules, but there may nevertheless be important individual differences that determine who can apply these rules in challenging circumstances. Suggestive support for such a possibility can be found in research on selfregulation. For example, White participants who have high levels of implicit prejudice show self-regulatory fatigue after engaging in an interracial interaction, which suggests that they may be inhibiting stereotypic thoughts during the interaction (Richeson & Shelton, 2003; see also Richeson et al., 2003). Additionally, selfregulatory fatigue has been found to reduce accommodative responses to hypothetical negative behaviors by romantic partners (Finkel & Campbell, 2001); this result suggests that inhibitory processes may be involved in forgiving partners’ transgressions. The goal of the current research was to extend these findings and examine directly the role of inhibition in social behavior. Inhibition was assessed via the Stroop task. This task requires participants to name the ink color in which each of a series of color words is written. When most of the color words are inconsistentwith the color of their ink, efficientnaming of the inkcolors requires inhibition of the meanings of the words (activated auAddress correspondence to William von Hippel, School of Psychology, University of New South Wales, Sydney 2052, Australia; e-mail: [email protected]. PSYCHOLOGICAL SCIENCE Volume 16—Number 7 497 Copyright r 2005 American Psychological Society tomatically by proficient readers). Therefore, the Stroop task provides a useful indicator of individual differences in inhibitory ability, and it is widely used for that purpose (e.g., Alain & West, 2000; Long & Prat, 2002). An important problem with most measures of inhibition, however, is that they contain both activation and inhibition components. A person who shows substantial slowing in color naming when presented with incongruent color words might have poor ability to inhibit word meaning. Alternatively, such a person might have a strong tendency to activate word meaning, so that greater inhibition would be required for that person to perform at the same level as others who do not have such a strong tendency for activation. Thus, if the Stroop task is to be interpreted as a measure of inhibitory ability, it is necessary to use it in a context that rules out activation differences as a likely cause of effects that emerge. In the current experiment, such a context was created by studying verbal and nonverbal blurting (cf. Swann & Rentfrow, 2001). Participants were asked to eat an unusual food item that was graphically similar to the animal part from which it had originated. Specifically, they were presented with a chicken foot cooked in a Chinese style that preserved the entire foot intact, including the claws. Social pressure not to blurt was manipulated by the way that the chicken foot was presented—by a Chinese experimenter who suggested prior to presentation that it was a culturally significant dish and a personal favorite or by a White experimenter who simply noted that it was Chinese food. This manipulation of social pressure made it possible to test whether the inhibitory component, rather than the activation component, of the Stroop task predicts blurting. If the activation component predicts blurting, then Stroop performance should have predicted blurting equally whether the experimenter was White or Chinese, as these two conditions involved equivalent activation of the unappealing aspects of the chicken foot. In contrast, if the inhibition component predicts blurting, then Stroop performance should have predicted blurting only when the experimenter was Chinese, as only in this condition were participants motivated to suppress a negative response. In addition to this manipulation of social pressure, the current experiment included a manipulation of divided attention. The goal of this manipulation was to provide sufficient distraction so that even good inhibitors would unintentionally blurt a negative response when confronted with the chicken foot. Again, however, divided attention was expected to affect blurting only when the experimenter was Chinese, as only with a Chinese experimenter would participants try to suppress a negative response. Thus, divided attention wasexpected to lead to increased blurting only with theChinese experimenter, andStroop performancewas expected to predict blurting only with the Chinese experimenter when participants had full attention available to devote to social control. Because nonverbal responses are more difficult to control than verbal responses (DePaulo, 1992), it seemed possible that nonverbal responses would be more sensitive to the experimental manipulations than verbal responses would be, and that inhibitory ability would more clearly predict nonverbal than verbal behavior. Nevertheless, previous research has demonstrated clear individual differences in verbal blurting (Swann & Rentfrow, 2001), and thus both nonverbal and verbal behavior were expected to be sensitive to inhibitory ability and social pressure.