Larks comments on Preference Inversion