Groundwork, Part I
A good will -- a rational actor, free to choose its actions, that always acts in the best way.
As with Aristotle, there's an understanding that talents and temperaments can be used or not for the right purpose, in the right way, at the right time -- and so they are not purely good. This matches with practical wisdom in the Nicomachean Ethics.
If something is to be purely good, that means first it must be pure. If it can be mistaken, it is not pure. Reason is the only thing that can be pure, and as such, it is the only thing that can possibly be the source of a good will.
Is Kant's criticism of the virtues that the ancients praised valid, when it comes to Aristotle? Did Aristotle say that each of the virtues could be good without limitation, or did they require something else? Is Aristotle's "in the right way, at the right time, etc." corresponding to Kant's good will? Maybe the difference is Kant's insistence on reason being the only way to have purely good will. I don't think Aristotle insisted on it, though he certainly valued reason for the sake of consistency, contemplation, judgment, understanding, practical wisdom.
"There is, however, something so strange in this idea of the absolute worth of a mere will, in the estimation of which no allowance is made for any usefulness, that, despite all the agreement even of common understanding with this idea, a suspicion must yet arise that its covert basis is perhaps mere high-flown fantasy and that we may have misunderstood the purpose of nature in assigning reason to our will as governor." (8) (4:394-5)
According to Kant, we have instinctual reactions that lead to our self-preservation and our happiness. The reason can only cause deviation from pure self-preservation, and therefore the reason must have some other purpose. This purpose is the determination of supreme laws, and apprehension of a priori principles. Reason was given to us as a practical, not just a contemplative faculty -- it influences the will, the capacity to act. Instinct can produce a will that is good as a means to achieving self-preservation and happiness. Only reason can produce a will that is good in itself, with regard to eternal, inviolable principles.
Are happiness and achievement of a good will so distinct from each other as Kant proposes? Do we not feel satisfaction from acting right, and in that case, does it not take part in happiness? In Part III, Kant says that we do feel inclined to be moral, that it does take part in our happiness as well as the good in itself.
We are divided selves, according to Kant -- selves of sense and selves of reason. This is how duty achieves such an important role -- we must submit our selves of sense, which only try achieve good empirically, and thus in flawed ways, to the reason, which can apprehend the supreme principle of morality. We may have inclinations, temperaments to the contrary, and so we must ignore them in favor of the law of reason. We do it out of a duty to that which is good in itself. The holy will, which always does what is good in itself and has no bad inclinations, needs no concept of duty -- it is a good will nonetheless.
For an action to have moral worth, it must have been contemplated by reason and decided on in accordance with the supreme principle -- only in this way is it done from duty, and not merely in conformity with duty -- because if it was done in conformity, but not from a sense of duty, it is easily changeable.
Also, in order for an action to be good, it must be purely decided. Since the ends achieved are never purely known, then the ends can't be the deciding factor that make some action good. Rather, it is in the intent itself -- and the only way an intent can be pure is if it is done in conformity with a maxim that is pure.
"Now, an action from duty is to put aside entirely the influence of inclination and with it every object of the will; hence there is left for the will nothing that could determine it except objectively the law and subjectively pure respect for this practical law, and so the maxim of complying with such a law even if it infringes upon all my inclinations." (14) (4:400-1)
A law free from all inclination would be universal. It would have to be known to apply in all circumstances. Therefore, to act in accordance with a law, and put aside one's inclinations, one would have to act according to a law that was universal. If the maxim for this action has not been determined, then we would have to look at the maxim and determine if it could be made into a universal law. If it could, then we should act that way. If it could not, it must not be according to a universal law, therefore it could have the taint of one's inclinations, and therefore it might not be good.
On what basis do we determine whether something should be a universal law, though? Why is it purely better not to cheat my customers? Even if it is determined to be purely better -- is good something that humans want or should try to act in accordance with? Kant gives an answer: "that it is an estimation of a worth that far outweighs any worth of what is recommended by inclination, and that the necessity of my action from pure respect for the practical law is what constitutes duty, to which every other motive must give way because it is the condition of a will good in itself, the worth of which surpasses all else." Kant is making the ungrounded statement that the good will is valued in itself, above all else. If this is what we value, then we should act this way. What could be questioned, then, is whether or not it is what we want. Or whether or not pure reason can connect with the world of sense. Or if reason can connect, can it be pure, and if not, is this all relevant.
Common human reason, so valuable in determining morality, still needs the strictness of practical philosophy because it can be easily seduced, and inclinations will be rationalized against the call of duty.