On the contrary, that’s why it’s perfect. There is so much evidence in so many different context that positive reinforcement yields better results than negative.

This is congruent with your donkey example.