Continuous-dual separation theorem for Hausdorff locally convex spaces (source code)

= Continuous-dual separation theorem for Hausdorff locally convex spaces
{title2=$X^*\text{ separates points of }X$}

= Continuous-dual point separation
{synonym}

The <continuous dual space> of a <Hausdorff> <locally convex space> separates points: for every nonzero $x$ there is a <continuous linear functional> $f$ with $f(x)\ne0$.

A separating family of continuous <seminorms> supplies $p(x)>0$. In the real case, define $g(tx)=tp(x)$ on the <span> of $x$ and extend it under $p$ by the <Hahn-Banach theorem>. The bound $|g(y)|\leq p(y)$ gives <continuity>. In a complex <vector space>, apply the real argument to the underlying real <vector space>, then use $f(y)=g(y)-ig(iy)$ to obtain a complex-linear <functional> with nonzero <real part> at $x$. The <Hausdorff> condition is essential: a nontrivial <vector> invisible to every continuous <seminorm> is also invisible to every <continuous linear functional>.