【內容提要】根據直到康德才得以完全確立的傳統道德責任觀念，一個人必須要爲之負責任的事情只是那
種直接依從他的意志作用力而完成的行爲，而獨立於意志的任何東西一直被認爲是與道德特性的歸屬毫無
關係的，這在不同的意義上既符合、也悖於道德直觀。然而，如果道德運氣是一個實在的道德概念，那麼
，傳統的道德責任觀念就必須加以修正，因爲我們也應該爲那種超越於人的控制能力（此根植於意志）的
事情負責。儘管如此，與首次嚴肅地將道德運氣概念引入當代道德理論論域的伯納德·威廉姆斯不同，本
文認爲，對於所謂“超越”必須有某些限制，而且這些限制可歸結爲本文所說的“道德關聯”也即與康德
式的意願性行爲的關聯，這種意願性行爲在本文中具體表現爲行爲者的過錯。這樣，道德運氣概念就有理
由繼續維繫於康德式的道德責任觀念，至少應該與後者保持某種連續性。
【關鍵詞】道德運氣 道德責任 意願性行爲 行爲者遺憾 正當性證明
Abstract: According to the traditional conception of moral responsibility, which up to Kant
had been established completely, what a person must be responsible for is nothing more than
what is done directly from her or his agency, whereas anything that is independent of the
will has been regarded as irrelevant to the ascription of moral status. This seems to be
both compatible and incompatible with moral intuition in different senses. If moral luck is
a real concept of morality, the conception has to be revised to the extent that we should be
held responsible for what is beyond the controlling capability grounded in the will.
Nevertheless, unlike Bernard Williams who first seriously introduced the concept of moral
luck into the contemporary realm of moral theory, the present paper maintains that there
must be some limits to "being beyond" and that they can come down to what the present paper
may call the “moral relations” to the Kantian voluntary action, which is concretely
manifested as agent’s（for example, the drivers’ or Gauguin’s）faults in the examples
conceived in the present paper. Thus understood, the concept of moral luck is justified in
keeping to the Kantian conception of responsibility, at least in maintaining some continuity
with the latter.
Key Words: Moral Luck Moral Responsibility Voluntary Action Agent-regret
Justification
當我們問道，一個人爲什麼要爲自己的行爲負責時，等待我們的一個簡潔而自明的答案就是，行爲者之所
以應該爲自己的行爲負責，恰恰是因爲行爲者要爲之負責的那個行爲是由行爲者自己選擇並完成的。這裏
提到“由行爲者自己選擇並完成的”本是回答問題的關鍵，但看上去似乎仍然含糊不明，因爲一個人除了
要爲他自己所做的行爲負責之外，在某些情況下，他還要爲那些並不是由他自己直接選擇並完成、但屬於
自己原則上能夠控制和管轄的事情負責，因此一個更爲嚴格的答案便是，行爲者之所以應該爲自己的行爲
負責，恰恰是因爲行爲者要爲之負責的那個行爲，要麼是由行爲者自己選擇並完成的，要麼是由他原則上
能夠加以控制的。行爲責任的歸屬最終是以行爲者的控制權能爲前提。這顯然意味着，只有對行爲者所能