2

Pardon me if I have the terminology wrong here, I'm not totally sure if this is what it's called in English.

I want to show that two sets are complementary in $\mathbb R^4$. From what I understand, this is equivalent to show that two sets are direct sum of $\mathbb{R}^4$. Is that true?

Here are my subspaces:

$H = \{(a,b,c,d)\in \mathbb{R}^4 \mid a + b + c + d = 0\}$

$\operatorname{span}((1,1,1,1))$

For starters, I proved that $H \cap \operatorname{span}((1,1,1,1))= \{(0,0,0,0)\}$

by saying that $\operatorname{span}((1,1,1,1))$ can be described as $\{(x,y,z,t)\in \mathbb{R}^4 \mid x = y = z = t\}$. Thus, $H \cap \operatorname{span}((1,1,1,1))= \{(x,y,z,t) \mid x + y + z + t = 0, x = y = z = t\}=\{(x,x,x,x)\mid 4x = 0\} = \{(0,0,0,0)\}$

Is this the right way to show this?

Now, to show that $H + \operatorname{span}((1,1,1,1)) = \mathbb{R}^4$:

I can say that $$ H + \operatorname{span}((1,1,1,1)) = \{(a,b,c,d)\in \mathbb{R}^4 \mid a + b + c + d = 0\} + \{(x,y,z,t)\in \mathbb{R}^4 \mid x = y = z = t\}. $$ I sort of have an idea of what I should be doing at this point, but I'm not too sure I'm even doing this properly.

How do I go about finishing my proof?

Furthermore, is showing that two subspaces are complementary equivalent to showing that their sum creates a basis?

Sorry for the poor notation. Any help is appreciated. Thank you.

egreg
  • 238,574
iaskdumbstuff
  • 815
  • 1
  • 6
  • 14
  • Hint: For any vector $(x,y,z,t)\in\mathbb{R}^4$, consider the vector $$(x',y',z',t'):=(x,y,z,t)-\frac{1}{4}(x+y+z+t)(1,1,1,1).$$Can you see why $(x',y',z',t')\in H$? – Frank Lu May 10 '18 at 21:44
  • Also, yes, you proved that $H\cap\text{span}{(1,1,1,1)}={(0,0,0,0)}$ correctly. – Just dropped in May 10 '18 at 22:02
  • https://math.stackexchange.com/questions/2143611/complementary-subspaces This question will answer your queries. – Sonali May 10 '18 at 22:06

1 Answers1

2

Yes, you have to prove that

  1. $H\cap\operatorname{span}((1,1,1,1))=\{(0,0,0,0)\}$
  2. $H+\operatorname{span}((1,1,1,1))=\mathbb{R}^4$

You can simplify the proof of the first fact by taking $a(1,1,1,1)=(a,a,a,a)\in\operatorname{span}((1,1,1,1))$ and imposing the condition that $(a,a,a,a)\in H$, that is, $$ a+a+a+a=0 $$ which of course implies $a=0$.

The second part can be done in several ways. One is to take $(x,y,z,w)\in\mathbb{R}^4$ and trying to express it in the form $$ (x,y,z,w)=(a,b,c,d)+r(1,1,1,1) $$ where $a+b+c+d=0$. This becomes a linear system $$ \begin{cases} a+r=x \\ b+r=y \\ c+r=z \\ d+r=w \\ a+b+c+d=0 \end{cases} $$ The matrix is $$ \left[\begin{array}{ccccc|c} 1 & 0 & 0 & 0 & 1 & x \\ 0 & 1 & 0 & 0 & 1 & y \\ 0 & 0 & 1 & 0 & 1 & z \\ 0 & 0 & 0 & 1 & 1 & w \\ 1 & 1 & 1 & 1 & 0 & 0 \end{array}\right] $$ and a simple elimination brings it in the form $$ \left[\begin{array}{ccccc|c} 1 & 0 & 0 & 0 & 1 & x \\ 0 & 1 & 0 & 0 & 1 & y \\ 0 & 0 & 1 & 0 & 1 & z \\ 0 & 0 & 0 & 1 & 1 & w \\ 0 & 0 & 0 & 0 & -4 & -x-y-z-w \end{array}\right] $$ which shows that the system has a (unique) solution.

Another way is to notice that $\dim H=3$ and $\dim\operatorname{span}((1,1,1,1))=1$. Since $H\cap\operatorname{span}((1,1,1,1))=\{(0,0,0,0)\}$, we know that $$ \dim\bigl(H+\operatorname{span}((1,1,1,1))\bigr)=3+1=4 $$ so the sum is $\mathbb{R}^4$. Why is $\dim H=3$? Because a basis is $$ \{(1,0,0,-1),(0,1,0,-1),(0,0,1,-1)\} $$

egreg
  • 238,574
  • Thanks a bunch for the help. I have a couple of follow up questions, if you don't mind answering them. First, I actually arrived at that system of equations, but without the a+b+c+d = 0; was that line absolutely necessary? It was the reason for my confusion, as I wasn't sure where to put that restriction.

    Secondly, why do we have to make sure that the intesection of both subspaces be only the zero vector to know that the dim of the sum is 4?

    Finally, is there a difference between a vector space being the direct sum of two subspaces and saying that two subspaces are complementary?

    – iaskdumbstuff May 10 '18 at 22:47
  • 1
    @user494405 The condition $a+b+c+d=0$ is necessary, because it ensures that $(a,b,c,d)\in H$. The Grassmann dimension formula is $\dim(H+K)=\dim H+\dim K-\dim(H\cap K)$, so the intersection being trivial is necessary for the dimension of the sum being the sum of the dimensions. Finally, yes: two spaces are complementary if their intersection is trivial and their (direct) sum is the whole space. – egreg May 10 '18 at 22:54