Action Alters Object Identification

[flocondavoine]

Article In Press, Journal of Experimental Psychology : Human Perception and Performance.

Action Alters Object Identification :
Wielding a Gun Increases the Bias to See Guns


Introduction

Stereotypes, expectations, and emotions influence an observer’s ability to detect and categorize objects as guns. In light of recent work in action-perception interactions, however, there is another, unexplored, factor that may be critical: The action choices available to the perceiver. In five experiments, participants determined whether another person was holding a gun or a neutral object. Critically, the participant did this while holding and responding with either a gun or a neutral object. Responding with a gun biased observers to report “gun present” more than did responding with a ball. Thus, by virtue of affording a perceiver the opportunity to use a gun, he or she was more likely to classify objects in a scene as a gun and, as a result, to engage in threat-induced behavior (raising a firearm to shoot). In addition to theoretical implications for event perception and object identification, these findings have practical implications for law enforcement and public safety.

Conclusion

The familiar saying goes that when you hold a hammer, everything looks like a nail. The apparent harmlessness of this expression fades when one considers what happens when a person holds a gun. We have shown here that, having the opportunity to use a gun, a perceiver is more likely to classify objects held by others as guns and, as a result, to engage in threat-induced behavior (in this case, raising a firearm to shoot).

What mechanism gives rise to this bias? One possibility is that it arises from either perceptual or conceptual priming. According to this account, holding a gun could lead observers to adopt particular expectations regarding the presence of firearms. For example, just as a stereotypes can evoke a bias to report “gun present”, so might the mere presence of a gun in the environment. This explanation, however, is not supported by our data. While using a gun to respond to the stimuli increased participants’ bias to see guns, the conspicuous presence a real gun that was never used did not alter the bias to detect guns. Thus, the presence of a gun did not evoke additional priming above and beyond the images and nature of the task, suggesting that action is critical to perceptual bias.

In the introduction, we outlined two separate means by which action may alter object identification. The first was the theory of event coding which posits that both perceptual and action-based representations arise from a common code. As such, planning an action with a given object or tool should bias observers to identify similar objects, a contention supported by our data. Just as planning an action that involves a gun influenced the perceptual detection of the presence of other guns, planning an action with a shoe influenced perceptual detection of shoes. What of the action-specific account of perception? This account argues that psychophysical judgments of dimensions such as distance, size, or speed are correlated with the perceiver’s ability to act. Hence, it is the modification of an observer’s action capabilities that is critical. The fact that both holding guns (tools which clearly change one’s capabilities when interacting with objects and people) and shoes (objects that do little to change one’s capabilities when not worn on the feet) altered perception is therefore less consistent with this hypothesis. At a minimum, this result challenges the possible extension of action-specific effects beyond perception of spatial properties to higher-order visual processes such as object identification.

While the current results fit best with the theory of event coding (Hommel et al., 2001), it is important to note two ways in which the present data refine and extend the theory’s claims. First, the current results extend the theory of event coding by demonstrating common codes influence object identification, not just feature detection. Prior to the experiments reported here, the theory of event coding has been tested in tasks where observers judge the orientation, direction, color, or spatial location of objects. Here, we suggest it is likely that the same mechanisms captured by such tasks also play a role in the determination of object identity. Second, our results demonstrate an effect of common codes on acting with and object, not just acting on an object. In previous research, responses were coded based on features such as spatial location (e.g. left or right) or the color of the button to be pressed. Here again, we have documented that common coding effects occur not just for the feature of the response, but also for objects used to make a response (see also Miles & Proctor, 2011). Together, our findings suggest that common coding representations are pervasive, and include many – if not all – types of actions, and influence many – if not all – aspects of perception.

The practical implications of the current results are also clear. It is true that the action-induced biases we observed were not specific to guns. That said, while the bias created by holding a shoe is benign, the act of wielding a firearm raises the likelihood that nonthreatening objects will be perceived as threats. This bias can clearly be horrific for victims of accidental shootings. According to the American Civil Liberties Union, approximately 25% of all law enforcement shootings involve unarmed suspects and, although it is impossible to derive a precise number, it is certain that many similar accidental shootings occur among private citizens. It is therefore in the public’s interest to determine the factors that can lead to accidental shootings as well as measures to reduce the impact of these factors. While several factors including one’s beliefs and expectations have been previously identified, the current results indicate that the mere act of wielding a firearm raises the likelihood that non-threatening objects will be perceived as threats. This bias is also detrimental for the armed officers and soldiers who act violently after mistakenly thinking they saw a gun. Public gun safety and police training courses should incorporate these findings into their training protocols.

Article en pièce jointe (pour les personnes connectées), aussi disponible ici : http://www.nd.edu/~jbrockm1/WittBrockmole_inPress_JEPHPP.pdf

[edit]Ajout conclusion