On the 'semi-empirical' nature of Absolute Differential Calculus