![\bf \textit{we know the range for }cos\left( (1)/(x) \right)\textit{ is }[-1,1]\textit{ therefore} \\\\\\ -1~\ \textless \ ~cos\left( (1)/(x) \right)~\ \textless \ ~1\impliedby \textit{multiplying all sides by }x^2 \\\\\\ -1x^2~\ \textless \ ~x^2cos\left( (1)/(x) \right)~\ \textless \ ~1x^2\implies -x^2~\ \textless \ ~x^2cos\left( (1)/(x) \right)~\ \textless \ ~x^2](https://img.qammunity.org/2018/formulas/mathematics/high-school/756nc7fpcz4oaprgzklz275nzvk1yxhlq2.png)
if the limit of -x² goes to "something", and the limit of x² goes to the same "something", if their limit coincide, and yet they're bounding the cosine expression, therefore, since the cosine expression is "sandwiched" between -x² and x², then the cosine expression "squeezes in" that little sliver between both -x² and x², and will inevitably go to the same limit.
