Illusions and models: a reply to Barrouillet and Lecas