magazine issue 6 2011 / Issue 12

This notion builds upon the idea that knowledge in memory is stored in an associative network (Anderson, 1983; Collins & Loftus, 1975; see Figure 1 for a simplified schematic depiction). Memory content that is closely related is stored close to each other. If one piece of content is activated, its activation spreads across the network and activates other related content. For example, reading or thinking about the word “harm” activates other related words such as “hit” or “fist”. More importantly, the activation of such content increases the likelihood that respective behavior will be exerted. Thus, thinking about aggression will increase the likelihood that someone shows aggressive behavior – especially in ambiguous situations, in which accessible thoughts can influence the interpretation of the situation. For instance, if someone bumps into another person, one might interpret this as a voluntary provocation if aggressive thoughts are accessible in mind. However, without such thoughts in mind, one might interpret the same behavior as an act of unthoughtfulness. Similarly, acting aggressively will increase thoughts related to aggression, which might then in turn increase aggressive behavior (for a review see Todorov & Bargh, 2002). This reasoning is in line with the GAM: Acting aggressively or thinking of aggression increases the accessibility of aggressive thoughts.

However, our own research has repeatedly demonstrated that acting aggressively does not inevitably increase thoughts related to aggression. More specifically, we could show that imaginary acts of aggression (Denzler, Förster, & Liberman, 2009; Denzler, Förster, Liberman, Rozenman, 2010) or taking revenge (Gollwitzer & Denzler, 2009) can reduce thoughts that are related to aggression (see also Denzler, Häfner, & Förster, 2010). We showed that thoughts related to aggression are inhibited in memoryif the aggressive act fulfills a goal. Aggressive acts without goal-fulfillment, however, increase the accessibility of aggression. To give an example, if we have the goal to harm a provocateur, fulfilling this goal (through harming the provocateur) reduces the accessibility of aggressive thoughts. However, hitting a punching bag – which is an aggressive act that does, not fulfill the goal of harming the provocateur – does not decrease aggressive thoughts (see Denzler et al., 2009 and Bushman, 1999). Why is that the case? If we have the goal to harm someone, everything that is related to the goal might be highly active in memory: the goal itself, means to achieve the goal, etc.

After the goal has been fulfilled, however, this heightened accessibility of aggressive thoughts loses its functionality and moreover might interfere with other goals. Consequently, goal-related constructs are inhibited in memory (see also Förster, Liberman, & Higgins, 2009); we think less of aggression after aggression has fulfilled a goal. As we will discuss later, we also behave less aggressive after such goal-fulfillment. Notably, according to our model, no aggressive act á la Lorenz reduces aggression. In fact, aggression without goal-fulfillment (such as hitting a punching bag while having the goal of harming a provocateur) increases aggression (Denzler et al., 2009).