Automatic imitation of pro- and antisocial gestures: Is implicit social behavior censored?

Abstract

According to social reward theories, automatic imitation can be understood as a means to obtain positive social consequences. In line with this view, it has been shown that automatic imitation is modulated by contextual variables that constrain the positive outcomes of imitation. However, this work has largely neglected that many gestures have an inherent pro- or antisocial meaning. As a result of their meaning, antisocial gestures are considered taboo and should not be used in public. In three experiments, we show that automatic imitation of symbolic gestures is modulated by the social intent of these gestures. Experiment 1 (N = 37) revealed reduced automatic imitation of antisocial compared with prosocial gestures. Experiment 2 (N = 118) and Experiment 3 (N = 118) used a social priming procedure to show that this effect was stronger in a prosocial context than in an antisocial context. These findings were supported in a within-study meta-analysis using both frequentist and Bayesian statistics. Together, our results indicate that automatic imitation is regulated by internalized social norms that act as a stop signal when inappropriate actions are triggered.

Introduction

The propensity to imitate is well documented in psychological science (Chartrand and van Baaren, 2009, Heyes, 2011). That is, there is now strong evidence that individuals unintentionally imitate the actions (Brass et al., 2000, Chartrand and Bargh, 1999, Genschow et al., 2013), postures (Scheflen, 1964), facial expressions (Dimberg, Thunberg, & Elmehed, 2000), and speech patterns (Bock, 1986) of the people they interact with. According to ideomotor theory, these imitative tendencies are the result of shared perception-action representations in the brain (Brass et al., 2000). In particular, it is assumed that the visual image of an action is part of its motor representation and therefore that observed actions trigger an automatic imitative response (Brass et al., 2000). Interestingly, imitation is known to have important social benefits in the sense that it smooths social interaction (Chartrand & Bargh, 1999), increases prosocial behavior (van Baaren, Holland, Kawakami, & van Knippenberg, 2004), and facilitates empathy (De Coster, Verschuere, Goubert, Tsakiris, & Brass, 2013). As a result, it has been suggested that automatic imitation is an inherently social phenomenon (van Baaren, Janssen, Chartrand, & Dijksterhuis, 2009) that is driven by the social profits it generates (Lakin et al., 2003, Stel et al., 2016, Wang and Hamilton, 2012). Specifically, it has been argued that operant conditioning causes individuals to associate imitative responses with social reward, which in turn leads them to use imitation, be it conscious or subconscious, as a means to facilitate social interaction (Lakin et al., 2003, Stel et al., 2016, Wang and Hamilton, 2012).

However, the potential of imitation to produce social profit is constrained by contextual bounds. As a result, social reward theories predict that imitative tendencies should be reduced when the context makes it unlikely that they will lead to positive social outcomes (Stel et al., 2016, Wang and Hamilton, 2012). For example, it is doubtful that imitation will facilitate social interaction when the imitated person is not looking at the imitator (Wang & Hamilton, 2012). From a social reward perspective, the absence of eye contact should thus result in weaker automatic imitation, which has now been confirmed across multiple studies (Forbes et al., 2016, Wang and Hamilton, 2014, Wang et al., 2011). Moreover, if there are no constraints in the environment, social reward theories predict that automatic imitation should depend on the individual’s motivation to interact with others. In support, research has reported increased imitative tendencies in situations that promote affiliation (Butler et al., 2016, Genschow and Schindler, 2016, Lakin and Chartrand, 2003, Rauchbauer et al., 2015), stimulate prosocial attitudes (Leighton et al., 2010, Wang and Hamilton, 2013), or signal threat (Grecucci et al., 2011, Rauchbauer et al., 2016, Rauchbauer et al., 2015), but decreased imitative tendencies in situations that emphasize self-focus (Hogeveen and Obhi, 2011, Spengler et al., 2010).

In sum, previous research suggests that automatic imitation is stronger when it has the potential to generate social reward (Lakin et al., 2003, Stel et al., 2016, Wang and Hamilton, 2012). However, does this mean that individuals imitate all behavior indiscriminately as long as they are motivated to affiliate and are in a context that allows affiliation? This is important because gestures in daily life are often used in a symbolic manner as a shorthand for social communication (Morris, 1994). For example, an upwards extension of the thumb can be used to demonstrate approval (i.e., thumbs up) and an upwards extension of the middle finger can be used to communicate insult (i.e., middle finger). Nevertheless, even though communicative gestures are highly prevalent in everyday life, most automatic imitation studies have instead focused on simple actions without a symbolic meaning (Brass et al., 2001, Brass et al., 2000, Catmur and Heyes, 2011, Heyes, 2011).

As an important exception, a single study looked at automatic imitation of communicative and non-communicative gestures performed by either a human or nonhuman agent (Liepelt, Prinz, & Brass, 2010). The results revealed less imitation of the nonhuman agent compared with the human agent when a communicative action was performed (e.g., peace sign) but not when a non-communicative action was performed (e.g., grasping). However, albeit interesting, this study was silent on whether imitative tendencies depend on the social message expressed by the observed gesture. If automatic imitation is driven by its social consequences, as previous research suggests (Stel et al., 2016, Wang and Hamilton, 2012), then it should be modulated by the degree to which an imitative response would be inappropriate according to the dominant social norm (Cialdini, Reno, & Kallgren, 1990). In particular, the social norm dictates that offensive gestures ought not to be used in public. As a result, if imitation would lead to the execution of a taboo gesture, then it should be inhibited to prevent norm violation.

In line with this hypothesis, research on word production has found an increase in response inhibition when participants are at risk of making a taboo error (Severens et al., 2011, Severens et al., 2012), suggesting that individuals monitor their behavior in order to inhibit the execution of actions that would otherwise violate the established norms. However, most research on social norms conducted hitherto has looked at language production (Dhooge and Hartsuiker, 2011, Severens et al., 2011, Severens et al., 2012) or intentional social conduct such as littering (Cialdini et al., 1990) and environmental conservation (Goldstein et al., 2008, Scheibehenne et al., 2016). As a result, it remains to be understood whether more spontaneous social behavior such as automatic imitation is regulated by social norms as well. In particular, if this is the case, then the mere observation of a taboo gesture should be sufficient to trigger an implicit stop signal that prevents the unintended execution of that gesture.

To examine the role of social norms in automatic imitation, the current study compared automatic imitation of pro- and antisocial gestures. More specifically, participants performed a prosocial (i.e., thumbs up) or antisocial (i.e., middle finger) gesture in response to a symbolic cue (i.e., M or D) while a hand on the screen performed either a congruent or incongruent gesture. Automatic imitation in this paradigm can be operationalized as a congruency effect with slower responses on incongruent trials than on congruent trials (Brass et al., 2000, Heyes, 2011). If automatic imitation is sensitive to social norms, then the congruency effect should be weaker when an antisocial gesture is observed than when a prosocial gesture is observed. Moreover, as semantic processes are known to be slow (Meyer et al., 2013, Özyürek, 2014), it can be expected that this difference will become larger when the delay between the presentation of the gesture and the presentation of the cue increases. The reason for this is that longer delays provide participants with more time to process the observed gesture before a response has to be formed.