On the conditions to be filled by a theory of life