The 'coping' hypothesis of stereotypic behaviour: a reply to Rushen