With no explanation, label text_A→text_B with either "strengthener" or "weakener".
text_A: PersonX turns PersonY's face PersonX then inspects
text_B: PersonX turns her child's face toward the camera.
weakener.