本文引介經濟學者法蘭克 (Frank, 1988) 道德情緒 (Moral Sentiments) 理論的兩個模型：以演化賽局為基礎的「承諾模型」與以心理學的配合法則 (Matching Law) 為要素的「名譽模型」，並且運用這兩個模型來討論德行 (Virtue) 層面的道德教育，及其在社會發展上所扮演的角色與價值。首先，道德教育在承諾模型的演化賽局中，被定義為：借由增强內在德性，人為的而且短期的改變合作者 (好人) 與背叛者 (壞人) 的人數比例，使合作者的人數增加。本文由模型中推導得出：一個社會由低度道德進步至高度道德的過程中，好人與壞人的平均獲利和社會整體的總利益皆會增加，不過此時好人的平均獲利會由大於壞人反轉成小於壞人。道德教育能帶來社會利益卻不能帶來社會正義。其次，本文接續名譽模型的論點推導得出，在道德教育能增強「名譽傳訊現象」時，即當名譽成了內在道德情緒的可靠外在「傳訊」特徵時，好人的比例在長期中會「自然」增加，而且社會整體的總利益也將增加。最後，本文提出「道德教育的內容不限於只在教道德」的兩點技術層面的意見供大家思考。

This study introduces Robert H. Frank’s moral sentiments theory (1988), and uses it to discuss the role of moral education. Firstly, following Frank’s commitment model and its evolutionary game process, moral education is defined as temporarily strengthening human inner moral emotion to increase the number of cooperator (good behavior) in a society. Two conclusions are found. (1) In a low moral society, the defector’s benefit is smaller than cooperator’s. In contrast, in a high moral society, the defector’s benefit is larger than cooperator’s. (2) Moral education (improving a society’s moral level) does not bring social justice. However, it does bring social benefit. Secondly, following Frank’s reputation model, we argue that because reputation can be a strong signal of individual’s inner virtue, moral education can strengthen individual’s reputation and increase the number of cooperator in a society. In the end, the study suggests that moral education does not need limitation in teaching morality.