PolarPlot[r[t], {t, 0, 2 Pi}] is basically the same as ParametricPlot[r[t] {Cos[t], Sin[t]}, {t, 0, 2 Pi}]. When r[t]<0 the point is placed in the opposite direction as the one you would expect from the value of the angle t. Try this:

Manipulate[
 PolarPlot[-2 t, {t, 0, 2 Pi}, 
  Epilog -> {PointSize[Large], Point[-2 s*AngleVector[s]], 
    Arrow[{{0, 0}, Pi/2 AngleVector[s]}]}],
 {{s, Pi/4}, 0, 2 Pi}]

This is a duplicate from the site https://mathematica.stackexchange.com/questions/185957/interpretation-of-polarplot

I will postpone my answer. Here we have an example of incorrect use of the polar radius function r[t], which by definition must be positive or zero, $r\ge 0$. Mathematica offers a kind of solution for $r<0$ using the Euler formula $-1=e^{i\pi}$, and interpreting multiplication by $-1$ as a rotation of $\pi$. In this case, $\pi$ is added to the argument t. This is where the strange data comes from.

PolarPlot[{-2*t, 2*t} , {t, 0, Pi}, AxesLabel -> {x, y}, 
 PlotLegends -> "Expressions"]