Summary
In a previous article (Kent, 1978) it was suggested that mistakes in mathematics should be viewed not simply as things to be corrected but as a source of learning about the thought processes of others. This paper comprises a further sequence of examples of the precise nature of mistakes made by pupils in learning mathematics, again within the general framework of an argument that in order to improve children's performance the teacher must become a sort of researcher, analysing the false, quasi‐logical structures which waylay the children at particular stages and, where possible, making appropriate generalizations.