Toward a Democratic Science