The claim that ‘learning how to learn’ is the central ability required for young people to be effective ‘lifelong learners’ is examined for various plausible interpretations. It is vacuous if taken to mean that we need to acquire a capacity to learn, since we necessarily have this if we are to learn anything. The claim that it is a specific ability is then looked at. Once again, if we acquire an ability to learn we do not need the ability to learn how to learn. After noting the implausibility of any such general ability, the paper goes on to examine the claim that certain specific but transferable abilities might satisfy the description ‘learning how to learn’. Various candidates are considered: forming and testing hypotheses and abduction are two promising ones, but each has significant weaknesses. Numeracy and literacy are thought to be more promising, but achievements at the national level leave a lot to be desired, despite the clear advantages for learning of being able to read, write and count. If we needed to learn how to learn before we learned how to read, write and count, it is unlikely that we would get anywhere. Finally, certain non‐cognitive dispositions and character traits rather than cognitive attributes are considered and, drawing on the work of Robert Dearden and others, it is suggested that the development of these aretaic (virtue‐based) and personal qualities rather than cognitive ones may be most decisive for developing independent learning in a range of subject matters.