Let $m,n\in \mathbb{N}$ and let $V, W$ be $\ \mathbb{R}$-vector spaces with $\dim V=n$ and $\dim W=m$.

Let $f:V\rightarrow W$ be a linear map.

I want to prove or disprove the following:

There is a linear map $f:V\rightarrow W$ with the following properties:

  1. $f$ is injective, $n=3$, $m=4$.
  2. $\text{rang}(f)=6$, $\text{Nullity}(f)=5$, $m=11$.
  3. $f$ is surjective, $n=4$, $m=3$.
  4. $f$ is injective, $n=2$, $\text{rang}(f)=3$.
  5. $\text{Nullity}(f)=0$, $n=3$, $m=5$.


I have done the following:

  1. Since $f$ is injective, then $\text{ker}(f) = \{0\}$ and so $\dim\text{ker}(f) = 0$.

    It holds that $\text{Im}(f)\subseteq W$. Therefore $\dim \text{Im}(f)\leq \dim W$.

    From the Rank–nullity theorem we have that \begin{equation*}\dim V = \dim \text{ker}(f) + \dim \text{Im}(f)=\dim \text{Im}(f)\leq \dim W\end{equation*}

    Therefore it cannot be that $n=4$ and $m=3$, right?

  2. From the Rank–nullity theorem we have that \begin{equation*}\dim V=\text{Defekt}(f)+\text{Rang}(f)=5+6=11 \Rightarrow n=11\end{equation*} How do we use the information of $m$ ?

  3. Since $f$ is surjective, it holds that $\text{Rang}(f)=W$.

    From the Rank–nullity theorem we have that \begin{equation*}\text{Rang}(f)=\dim V-\text{Nullity}(f)\leq \dim V \Rightarrow \dim W\leq \dim V \Rightarrow m\leq n\end{equation*}

    Therefore there can be a linear surjective map with $n=4$ and $m=3$, right?

  4. Since $f$ is injective, it holds that $\text{ker}(f) = \{0\}$ and so $\dim\text{ker}(f) = 0$.

    It holds that $\text{Nullity}(f)=\dim \text{ker}(f)=0$.

    From the Rank–nullity theorem we have that \begin{equation*}\text{Rang}(f)=\dim V-\text{Nullity}(f) \Rightarrow 3=4-0 \Rightarrow 3=4\end{equation*} A contradiction.

    Therefore there cannot be a linear surjective map with $n=2$ and $\text{Rang}(f)=3$, right?

  5. Do we apply here the Rank–nullity theorem? But how?


For 1., you can consider $f\colon \mathbb R^3\to\mathbb R^4$ given by $$ f(x,y,z)=(x,y,z,0) $$ which is injective.

For 2., you can consider $f\colon \mathbb R^{11}\to\mathbb R^{11}$ given by $$ f(x_1,\ldots,x_{11})=(x_1,\ldots,x_6,0,\ldots,0), $$ since $\ker f$ equals the set of vectors of the form $(0,\ldots,0,x_7,\ldots,x_{11})$, which is $5$-dimensional.

For 3., you can consider $f\colon \mathbb R^4\to\mathbb R^3$ given by $$ f(x_1,\ldots,x_4)=(x_1,x_2,x_3), $$ which is surjective.

For 4., you can show it is impossible regardless of the value of $m$ - more generally, $\dim f(V)\leq \dim V$ regardless of the linear map $f$.

For 5., you can consider $f\colon \mathbb R^3\to\mathbb R^5$ given by $$ f(x_1,\ldots,x_3)=(x_1,x_2,x_3,0,0), $$ which is injective.

Notice my examples in 1., 3., and 5. had exactly the same form, namely $$ f\colon \mathbb R^n\to\mathbb R^m,\qquad f(x_1,\ldots,x_n)=(x_1,\ldots,x_{\min(m,n)},0,\ldots,0). $$ You can see that $\ker f$ consists of all elements of $\mathbb R^n$ whose first $\min(m,n)$ entries are equal to zero, so $$ \dim\ker f=n-\min(m,n). $$ Similarly, $$ \dim f(\mathbb R^n)=\min(m,n). $$ The rank-nullity theorem holds because $$ \bigl(n-\min(m,n)\bigr)+\min(m,n)=n. $$ Moreover $f$ is injective iff $\dim\ker f=0$, i.e. $n=\min(m,n)$ which is equivalent to $n\leq m$. And $f$ is surjective iff $\dim f(\mathbb R^n)=m$, i.e. $\min(m,n)=m$ or equivalently $m\leq n$. Consequently $f$ is bijective if and only if $m=n$.

| cite | improve this answer | |
  • $\begingroup$ At your last sentence you have that f is injective iff n<m. Therefore isn't my proof of 1 correct and there is not such a map? $\endgroup$ – Mary Star Feb 2 at 8:57
  • $\begingroup$ You wrote $n=3$ and $m=4$ in the question, but then later you wrote $n=4$ and $m=3$. I used the first one. And just to be clear, it is not true that $f$ is injective iff $n<m$. It is important to say $n\leq m$ instead of $n<m$. $\endgroup$ – pre-kidney Feb 2 at 8:58
  • $\begingroup$ Ahh ok!! Could you explain to me the proposition 2? Why does this f satisfy the desired conditions? $\endgroup$ – Mary Star Feb 2 at 10:51
  • $\begingroup$ What part of 2. doesn't make sense? By definition, the nullity equals $\dim \ker f$. I explained why this equals $5$. Also, what you are calling rang($f$) is the same as $\dim f(V)$, or in words the dimension of the image of $f$. From the form of the function $f$ I wrote, you can see that the image of $f$ is $6$-dimensional, corresponding to the first $6$ entries of the $11$-tuple. Since I have given what I believe to be a complete explanation already, if something doesn't make sense to you the onus is on you to ask a concrete and specific question that I can respond to. $\endgroup$ – pre-kidney Feb 2 at 22:48

Since these are linear maps between finite dimensional vector spaces $V_n \longrightarrow W_m$ so each of them (if it exists) can be represented by a $m \times n$ matrix $A$.

For (1): Here $A$ must be $4 \times 3$ matrix. For injective map, no free columns in $A$, so take $$A_{4 \times 3}=\begin{bmatrix}1&0&0\\0&1&0\\0&0&1\\0&0&0\end{bmatrix}$$

For (2): $\text{dim} V=\text{rank}+\text{nullity}=6+5=11$, so $A$ should have $6$ pivot columns and $5$ free columns. For example, $$A_{11 \times 11}=\begin{bmatrix}\mathbf{e_1} & \mathbf{e_2} & \ldots & \mathbf{e_6} & \mathbf{0} & \ldots & \mathbf{0}\end{bmatrix}.$$

For (3): Here $A_{3 \times 4}$. So max rank of $A$ is $3$. For surjective map, we need full rank. Thus $$A_{3 \times 4}=\begin{bmatrix}1&0&0&0\\0&1&0&0\\0&0&1&0\end{bmatrix}$$

For (4): Here $A_{m \times 2}$. $\text{rank}=3$, so $m \geq 3$. For injective map $A$ should not have free columns. But maximum rank of this matrix can be $2$ and $m$ being at least three, there will always be a free column. So not possible.

For (5): Here $A_{5 \times 3}$. Nullity $0$ means injective map, so no free columns in $A$. Maximum rank of this matrix can be $3$. Thus $$A_{5 \times 3}=\begin{bmatrix}1&0&0\\0&1&0\\0&0&1\\0&0&0\\0&0&0\end{bmatrix}$$

| cite | improve this answer | |

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.