Evidence that individuals distinguish between moral and conventional rules is reviewed. Moral rules prohibit actions that result in victims (e.g., violence, stealing, etc.). Conventional rules prohibit actions that do not result in victims (e.g., not saying please, dressing in opposite sex clothes). Previous theoretical accounts of the development of the moral/conventional distinction are discussed. These theories are contrasted with an approach that is developed here. It is proposed that there is a mechanism, a Violence Inhibition Mechanism (VIM), that is responsible for the previously observed aversive arousal response to the distress of others. It is proposed that this aversive arousal response is a prerequisite for the development of the moral/ conventional distinction, the moral emotions and the inhibition of violent behaviour. Previous accounts have stressed role taking as a prerequisite for the moral/ conventional distinction. However, this was found not to be the case. Autistics, already known to be lacking a 'Theory of Mind' and therefore unable to role take, were found to make the moral/ conventional distinction. It was hypothesized that Antisocial Personality Disorder (APD) might be a consequence of a lack of VIM; the clinical description of APD stresses their lack of the moral emotions and their inability to inhibit their violent actions. In line with this, APD subjects were not found to make a moral/conventional distinction. Two rival explanations of APD were investigated: that APD is due to an inability to role take and; that APD is due to frontal lobe damage. APD subjects were not found to be impaired in either of these respects in comparison to criminal controls. A final investigation focused on the emotion attributions of APD subjects. It was hypothesized, given the contention that VIM is a prerequisite for the development of the moral emotions, that APD subjects might make anomalous attributions in victim situations though their attributions of other situations should prove normal. This study observed that while the attributions of APD subjects and criminal controls did not differ if the emotions attributed were happiness, sadness or embarrassment there was significant difference in victim situations where APD subjects were less likely to attribute guilt and more likely to attribute indifference than criminal controls. This finding was taken as indirect support of the VIM position. Additional tests, and implications of the VIM model are then discussed.

本論文嘗試以行為引發機制與行為抑制機制的觀點，探討非法且屬人際層面的男性攻擊行為，從中評估常表現攻擊行為個體的認知、情緒與行為缺陷。由於攻擊行為的特殊性，個體表現攻擊行為時，相關的行為機制包括表現該行為可能得到獎賞的行為引發機制﹔表現該行為可能得到懲罰的行為抑制機制﹔與攻擊他人時，被害者的非語文訊息引發的個體的暴力抑制機制。研究一根據暴力抑制機制的理論內容發展實驗工具，透過違反道德與違反慣例圖片組的呈現，了解高、中、低攻擊組的表現差異。結果發現高攻擊組對兩違反情境的嚴重程度評估較低，同時，在違反道德圖片組刺激下，高攻擊組表現較少的道德情緒與較少的同理行為。研究二則透過研究一發展的圖片組評估個體的暴力抑制機制運作能力，並結合Patterson與Newman的四階段理論發展實驗刺激，透過情境的操弄，了解高攻擊組之行為引發機制與行為抑制機制運作情形。結果發現經由兩套理論(暴力抑制機制與四階段理論)的結合，可將高攻擊組分為兩類，一類為暴力抑制機制運作正常，但行為引發機制運作過強﹔另一類為暴力抑制機制無法運作者。以認知、情緒與行為缺陷評估兩類高攻擊行為者發現，具有認知、情緒與行為缺陷的高攻擊行為者只有第二類。 / Based on a behavioral activation mechanism and two behavioral inhibition mechanisms, unlawful and interpersonal male aggression was studied. Moreover, the possibility of individuals’ cognitive, emotional, and behavioral deficiencies was evaluated. Due to the special quality of aggressive behavior, when a person acts aggressively, there will be three processes involved. They are the behavioral activation mechanism activated by reward, the behavioral inhibition mechanism induced by punishment, and the Violence Inhibition Mechanism (VIM) triggered by victims’ cues of distress. In order to study these three processes, two sets of pictures were developed from study 1 in order to evaluate how participants functioned their VIM. The results showed differences between High, Medium, and Low Aggression Group. The High Aggression Group evaluated incidents of moral/ conventional transgression less serious than the other two groups did. Also, they showed less moral emotions and less empathic behaviors than the other two groups did. Based on the sets of pictures developed from Study 1, a group of young offenders were divided into two groups based on their VIM functioning. Further, a survival game was developed from Patterson and Newman’s four-stage model. By manipulating situations to present reward/punishment, those young offenders showed different aggressive patterns. Young offenders with good VIM functioning behaved more aggressively when there were rewards. Relatively, young offenders with poor VIM functioning were less likely to be influenced by either reward or punishment. The implications were discussed in the article.