As children, we are often told both what to do and what to think. For a child to learn at all, it must in the first instance simply trust those, such as parents, who teach it things; and this goes for practical as well as theoretical learning. Doubting is necessarily something that comes later, for to be able to doubt one must have some beliefs already, e.g. concerning what sort of reasons count as good reasons, and what count as bad. But in growing up, a person does, or should, develop the capacity for rational doubt, and also the capacity for rational resistance to being told what to do. The first capacity constitutes a critical faculty, and the second is an essential constituent of practical autonomy.

A proponent of the value and importance of these two capacities may claim the following: firstly, that a grown-up person should only believe something if they have satisfied themselves that there are good reasons for believing it; secondly, that a grown-up person should only do something if they have satisfied themselves that there are good reasons for doing it. The second part of this claim clearly needs restricting, since one can do things just for the fun of it, such as hopping over a puddle, or absent-mindedly, such as tapping one's fingers, without thereby appearing less than fully rational. But if we are considering the case where someone is telling you, or asking you, to do something, mightn't it be true that autonomy requires you to obey, or accede to the request, only if you are satisfied that there is good reason to do so?