On the hypotheses which lie at the foundation of geometry