Consider a single particle moving around the origin in a circle. If the particle's coordinates are $(x,y)$ at some time, it will reach $(x,-y)$ after some time as they lie on the circle.

The position vector of the particle is rotating about the origin. Now rotation of a vector by an angle $Z$ clockwise is equivalent is to rotating the coordinate axes anticlockwise by $Z$.

The components in the new axes rotated anticlockwise are given by : $$ x' = x\cos(Z) + y\sin(Z)$$ $$ y' = -x\sin(Z) + y\cos(Z)$$

These are also the component transformations if I rotate the vector clockwise and keep my axes the same.

The question is : By rotation of the axes , we cannot transform the vector $(x,y)$ into $(x,-y)$ since there is no angle $Z$ that satisfies the equations for this pair ( x and y not 0 ). But on rotating the vector we get $(x,-y)$. This seems to be a paradox. What am I missing here?

  • $\begingroup$ Whatever $+Z$ works with one case, $-Z$ should also work on the other. A rotation about an angle $+Z$ is equivalent to a coordinate transformation about an angle $-Z$. $\endgroup$ Nov 8, 2017 at 20:39
  • $\begingroup$ $Z=\pi$ doesn't return (0,-1) for input (0,1)? $\endgroup$
    – Kyle Kanos
    Nov 8, 2017 at 20:39
  • $\begingroup$ @KyleKanos it doesnt return (2,-3) for (2,3) $\endgroup$
    – cobra121
    Nov 8, 2017 at 20:41
  • $\begingroup$ Try $Z\sim225.25^\circ$ for that case. $\endgroup$
    – Kyle Kanos
    Nov 8, 2017 at 20:46
  • 2
    $\begingroup$ I don't really see how you guys are getting a passive rotation that maps $(x,y)\rightarrow (x,-y)$. This is a reflection of the $y$ axis about the $x$ axis and therefore is not a rotation. $\endgroup$
    – Diracology
    Nov 8, 2017 at 21:29

1 Answer 1


Indeed you will not get the transformation $(x,y)\rightarrow(x,-y)$ by a rotation of the coordinate system (passive rotation) since this transformation actually corresponds to a reflection, namely $$S= \begin{pmatrix} 1&0\\0&-1 \end{pmatrix},\tag1$$ which does not belong to the group of rotations in the plane, $SO(2)$.

On the other hand, by an active rotation $R(Z)$, $$R(Z)= \begin{pmatrix} \cos Z&-\sin Z\\ \sin Z&\cos Z \end{pmatrix},\tag2 $$ of a definite angle $Z$ you definitely can obtain the transformation $(x,y)\rightarrow(x,-y)$ for some definite vector $(x,y)$. The paradox you mentioned could be restated as following: How can a reflection matrix (in the passive view) be identified with a rotation matrix (in the active view)? There is no paradox though, since (1) and (2) are not supposed to be identified as operators. For example, both matrices $$S= \begin{pmatrix} 1&0\\0&-1 \end{pmatrix}, \quad\mathrm{and}\quad R(3\pi/2)= \begin{pmatrix} 0&1\\ -1&0 \end{pmatrix} $$ take the vector $(1/\sqrt 2,1/\sqrt 2)$ to $(1/\sqrt 2,-1/\sqrt 2)$, but $S$ takes $(0,1)$ to $(0,-1)$ whereas $R$ maps $(0,1)$ to $(1,0)$. The explicit representation of the operators above already shows they are different.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.