Predispositions to Approach and Avoid Are Contextually Sensitive and Goal Dependent Susan Bamford and Robert Ward Bangor University The authors show that predispositions to approach and avoid do not consist simply of specific motor patterns but are more abstract functions that produce a desired environmental effect. It has been claimed that evaluating a visual stimulus as positive or negative evokes a specific motor response, extending the arm to negative stimuli, and contracting to positive stimuli. The authors showed that a large congruency effect (participants were faster to approach pleasant and avoid unpleasant stimuli, than to approach unpleasant and avoid pleasant stimuli) could be produced on a novel touchscreen paradigm (Experiment 1), and that the congruency effect could be reversed by spatial (Experiment 2) and nonspatial (Experi- ment 3) response effects. Thus, involuntary approach and avoid response activations are not fixed, but sensitive to context, and are specifically based on the desired goal. Keywords: approach, avoid, valence, evaluation, response The fundamental question facing any cognitive agent might be the decision whether to approach or avoid. It can be life- threatening to approach a dangerous object, yet equally maladap- tive to avoid beneficial ones. We might then expect cognitive agents to have sophisticated mechanisms for evaluating stimuli, and for initiating appropriate approach and avoid responses. It is therefore surprising that previous research suggests very unsophis- ticated approach and avoid responses. For example, the avoid response in a frog can consist simply of leaping to where it is darker (Lettvin, Maturana, McCulloch, & Pitts, 1959). Previous research has suggested that people likewise have pre- dispositions to respond in specific ways to valenced stimuli. Chen and Bargh (1999) found that when participants were asked to judge the valence (pleasant or unpleasant) of a word and then respond by using a lever, participants were faster to pull in response to pleasant words, and to push in response to unpleasant words. Chen and Bargh (1999; see also Solarz, 1960; Cacioppo, Priester, & Bernston, 1993; Forster & Strack, 1996) argued that people have a predisposition to approach positive items by pulling them closer, and to avoid negative ones by pushing them away. That is, the evaluation of the stimulus is automatically associated with a spe- cific muscle response: flexing a bicep for an approach movement, and extending the tricep for an avoid movement. However, at the level of the effectors, no single set of responses can be appropriate for every situation. In some cases when en- countering an unpleasant stimulus, like a spider, you would be unlikely to reach out and try to push it away. You would be more likely to jump back and flex your muscles away from it. Similarly, when encountering a pleasant object, like a pizza, the first response you need in order to get it is to extend your arm to reach it and pick it up; the flexing comes later. It seems as if the appropriate initial response, of flexing or extending to pleasant and unpleasant stim- uli, depends on situational demands. So if a pleasant stimulus predisposed us to flex our muscles and draw something toward us, then in a situation in which an extended reach was first necessary, the predisposition would have to be overridden. This would at least cause a delay and would defeat the value of automatic predispo- sitions to behavior. Here we ask whether approach and avoid responses in humans are highly specified, as argued elsewhere (Chen & Bargh, 1999; Solarz, 1960; Duckworth, Bargh, Garcia, & Chaiken, 2002), or whether situational factors exert a strong in- fluence on which behavior is activated. In many papers that investigate how the evaluation of valenced stimuli affects subsequent flex and extend movements, the authors admit that there may be situations in which context could influence or change the behaviors that are associated with approaching and avoiding valenced stimuli, even while arguing for highly specific response activations. For example, Chen and Bargh (1999) state “it may be possible to generate quite different effects within the same paradigm. . ..different social situations call for different re- sponses. . ..although the scope of the current experiments is not designed to address these issues, they are necessary avenues for future research.” Clore and Ortony (2000) also argued for a dis- sociation between appraisal and specific behaviors. And recently Rotteveel and Phaf (2004) showed that tendencies for arm flexion and extension are not automatic consequences of automatic affec- tive evaluation. They did however find a link between explicit evaluation and arm flexion and extension, and they discuss this with relation to the possible effect of situated meaning. They suggest that “if action tendencies for arm flexion and extension depend on conscious appraisals, the situated meaning and context for these movements would be incorporated into these processes.” Susan Bamford and Robert Ward, Wolfson Centre for Clinical and Cognitive Neuroscience, Bangor University. This research was supported by a Biotechnology and Biological Sci- ences Research Council (BBSRC) PhD studentship that was awarded to Susan Bamford. Correspondence concerning this article should be addressed to Susan Bamford, PhD, Wolfson Centre for Clinical and Cognitive Neuro- science, Bangor University, Bangor, UK, LL57 2AS. E-mail: s.bamford@bangor.ac.uk Emotion Copyright 2008 by the American Psychological Association 2008, Vol. 8, No. 2, 174 –183 1528-3542/08/$12.00 DOI: 10.1037/1528-3542.8.2.174 174