The epistemology of induction breaks down even before having taken its first step. It cannot start from sense data or perceptions and build our theories upon them, since there are no such things as sense data or perceptions which are not built upon theories (or expectations, that is, the biological predecessors of linguistically formulated theories).
#