For the solution set of the homogeneous equation $A\mathbf{x}=\mathbf{0}$, I am told I can visualize it as follows:

  • the single point $\mathbf{0}$, when $A\mathbf{x}=\mathbf{0}$ has only the trivial solution,
  • a line through $\mathbf{0}$, when $A\mathbf{x}=\mathbf{0}$ has one free variable,
  • a plane through $\mathbf{0}$, when $A\mathbf{x}=\mathbf{0}$ has two free variables. (For more than two free variables, also use a plane through $\mathbf{0}$.)

My question really concerns the last point. I feel as though I understand everything except for how more than two free variables corresponds to a plane through $\mathbf{0}$.

In general, using two variables, we have a line. Using three variables, we have a plane. Using four or more variables...what do we have then? For instance, the homogeneous system $$ x_1-3x_2-9x_3+5x_4=0\\[0.5em] x_2+2x_3-x_4=0 $$ row reduces as $$ \begin{bmatrix} 1 & -3 & -9 & 5 & 0\\ 0 & 1 & 2 & -1 & 0 \end{bmatrix} \sim \begin{bmatrix} 1 & 0 & -3 & 2 & 0\\ 0 & 1 & 2 & -1 & 0 \end{bmatrix} $$ to yield the solution set $$ \begin{bmatrix} x_1\\ x_2\\ x_3\\ x_4 \end{bmatrix} = \begin{bmatrix} 3x_3-3x_4\\ -2x_3+x_4\\ x_3\\ x_4 \end{bmatrix} =x_3 \begin{bmatrix} 3\\ -2\\ 1\\ 0 \end{bmatrix} +x_4 \begin{bmatrix} -2\\ 1\\ 0\\ 1 \end{bmatrix}. $$ Apparently an "appropriate geometric picture" for the solution set is a plane through the origin.

Can someone explain this? Why can the solution set be viewed as a plane through the origin? Why can one generally look at the solution set as a plane through the origin when there are two or more free variables? Is this entirely accurate?

  • 1
    $\begingroup$ Well it depends on what your definition of a plane is. Some people call any k dimensional linear subspace of $R^n$ a (higher dimensional) plane if $k\ge 2$, which is apparently what your source did. Then for some other people a plane has to be 2-dimensional, and higher dimensional analogues are called flats. In either case the solution space in your example is a plane in $R^4$. $\endgroup$ – Hyperplane Mar 15 '17 at 19:23
  • $\begingroup$ How fitting your username is...on the good ole Wiki I see that "For $n$ variables, each linear equation determines a hyperplane in $n$-dimensional space. The solution set is the intersection of these hyperplanes, which may be a flat of any dimension." Can't wrap my head around it too much, but I think that answers my question (wish the author had made an aside, since he actually does talk about hyperplanes near the end of the text). Thanks for the clarifying comment. $\endgroup$ – interrogative Mar 15 '17 at 19:30

A plane in a finite dimensional vector space can be described as a two dimensional subspace.

Note that, when a subspace $F$ is defined by a set of linear equations (after a basis has been chosen) in a vector space $E$ of dimension $n$, the codimension of the subspace, i.e. the number $$\operatorname{codim}_EF=\dim E-\dim F,$$ is equal to the rank of the system of equations, i.e. the rank of its associated matrix.

If the codimension is, say, $3$ (which means the subspace of solutions is defined by $3$ linearly independent equations), we have a subspace of dimension $n-3$: it will be a line if the big space $E$ has dimension $4$, a plane if it has dimension $5$, a ‘solid’ (isomorphic to $K^3$) if it has dimension $6$.


If you think that visualizing $\mathbb R^n$ is not difficult, then it is also not difficult to see that the solution set

$$\ker A:=\{\vec{x} \in \mathbb R^n \mid A\vec{x}=0\}$$

is a $k$ dimensional vector space, where $k=n-\mathrm{rank} A$. But since it is a linear subspace, it has a basis $e_1,..,e_k$. Send each of these to the standard basis in $\mathbb R^k$ and observe that we obtain an isomorphism. Moreover, if the basis you specified originally was orthonormal, we have an isometric isomorphism, so in a rigorous sense, everything is the same (including inner product etc.).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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