If the 'new problem' (unknowable thresholds) is epistemological rather than semantic, rejecting indu...