To accept subjectively that a model accounts for real world processes, hence external to science