The purpose of this experiment was to examine a causal relation between perspective-taking and verbal aggression. A participant conducted verbal communication through a computer with another supposed participant (actually, a computer program). Half of the participants first performed a task that was designed to activate perspective-taking, whereas the other half first performed a different task that was designed not to activate perspective-taking. In verbal communication, it was found that those who had not activated perspective-taking increased the number of instances of verbal aggression as the alleged counterpart became more aggressive, whereas those who had activated perspective-taking did not. This finding suggests that activation of perspective-taking has the effect of suppressing an increase in verbal aggression toward a person who increases verbal aggression.