It is because the quantity of attainable phrase sequences boosts, and also the styles that notify results turn out to be weaker. By weighting phrases within a nonlinear, distributed way, this model can "understand" to approximate terms and never be misled by any unfamiliar values. Its