Geometric Interpretation of Reversing the Components of a Vector

Metronome

Junior Member
Joined
Jun 12, 2018
Messages
103
Is there a nice geometric interpretation of reversing the components of a vector? So given some [imath]\begin{bmatrix}x_1 \\ x_2 \\ x_3\end{bmatrix}[/imath], I want to visualize how it relates to [imath]\begin{bmatrix}x_3 \\ x_2 \\ x_1\end{bmatrix}[/imath]. Is there a specific linear transformation that accomplishes this reversal? I have tried graphing a few vectors and their "algebraic inverses," and all I have noticed so far is that the length of the original vector is preserved.

Ultimately I am interested in a geometric interpretation of performing this action followed by taking a dot product with another vector (which in and of itself has a geometric interpretation of projection followed by scaling), as this is very close to a discrete convolution. Discrete convolution has its own geometric interpretation defined on functions of a discrete variable, but these are basically identical to vectors, so I was wondering what the operation would look like in geometric vector space.
 
Is there a nice geometric interpretation of reversing the components of a vector? So given some [imath]\begin{bmatrix}x_1 \\ x_2 \\ x_3\end{bmatrix}[/imath], I want to visualize how it relates to [imath]\begin{bmatrix}x_3 \\ x_2 \\ x_1\end{bmatrix}[/imath]. Is there a specific linear transformation that accomplishes this reversal? I have tried graphing a few vectors and their "algebraic inverses," and all I have noticed so far is that the length of the original vector is preserved.

Ultimately I am interested in a geometric interpretation of performing this action followed by taking a dot product with another vector (which in and of itself has a geometric interpretation of projection followed by scaling), as this is very close to a discrete convolution. Discrete convolution has its own geometric interpretation defined on functions of a discrete variable, but these are basically identical to vectors, so I was wondering what the operation would look like in geometric vector space.
Try doing the equivalent in two dimensions, and you should see that it's a reflection across the line y=x.

What reflection is it in three dimensions?
 
Try doing the equivalent in two dimensions, and you should see that it's a reflection across the line y=x.

What reflection is it in three dimensions?
It looks like this is a reflection around the span of the sum of the two vectors. [imath]\begin{bmatrix}x_1 \\ x_2 \end{bmatrix} + \begin{bmatrix}x_2 \\ x_1\end{bmatrix} = \begin{bmatrix}x_1 + x_2 \\ x_1 + x_2\end{bmatrix}[/imath], whose span is [imath]y = x[/imath] for any initial choice of vector. [imath]\begin{bmatrix}x_1 \\ x_2 \\ x_3\end{bmatrix} + \begin{bmatrix}x_3 \\ x_2 \\ x_1\end{bmatrix} = \begin{bmatrix}x_1 + x_3 \\ 2 x_2 \\ x_1 + x_3\end{bmatrix}[/imath], whose span is a different line depending on the initial choice of vector. So in three dimensions the transformation over the entire vector space is not a reflection?
 
It looks like this is a reflection around the span of the sum of the two vectors. [imath]\begin{bmatrix}x_1 \\ x_2 \end{bmatrix} + \begin{bmatrix}x_2 \\ x_1\end{bmatrix} = \begin{bmatrix}x_1 + x_2 \\ x_1 + x_2\end{bmatrix}[/imath], whose span is [imath]y = x[/imath] for any initial choice of vector. [imath]\begin{bmatrix}x_1 \\ x_2 \\ x_3\end{bmatrix} + \begin{bmatrix}x_3 \\ x_2 \\ x_1\end{bmatrix} = \begin{bmatrix}x_1 + x_3 \\ 2 x_2 \\ x_1 + x_3\end{bmatrix}[/imath], whose span is a different line depending on the initial choice of vector. So in three dimensions the transformation over the entire vector space is not a reflection?
Not quite. Keep thinking ... but more simply. Don't bias yourself by starting in a way that depends on the given vector!

In three dimensions, you can reflect across a plane as well as a line. If you want to continue your line of thinking, consider what plane always contains the sum of the vectors?
 
Not quite. Keep thinking ... but more simply. Don't bias yourself by starting in a way that depends on the given vector!

In three dimensions, you can reflect across a plane as well as a line. If you want to continue your line of thinking, consider what plane always contains the sum of the vectors?
Might you be switching some of the axes???
Think I got it, thanks! The action turns [imath]\hat i[/imath] into [imath]\hat k[/imath] and vice-versa, so the matrix is [imath]\begin{bmatrix}0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0\end{bmatrix}[/imath]. I found some manim code I could plug my matrix into and it looks like the plane of interest is [imath]z = x[/imath].
 
Top