ON THE NECESSITY OF INCLUDING THE OBSERVER IN PHYSICAL THEORY