On the classification of sciences