for many-valued deflection