Notes - Linear Algebra I MT22, Dimension Formula


Dimension formula

Can you state the full result of the dimension formula (aka Grassmann’s Identity) for $U, W, V$?


Let $U, W$ be subspaces of a finite dimensional vector space $V$. Then

\[\dim(U+W) + \dim(U \cap W) = \dim(U) + \dim(W)\]

The dimension formula states:

Let $U, W$ be subspaces of a finite dimensional vector space $V$. Then $\dim(U+W) + \dim(U \cap W) = \dim(U) + \dim(W)$.

If you let $U \cap W$ have a basis $\{v _ 1, \ldots, v _ m\}$, what bases does the Steinitz exchange lemma guarentee for $U$ and $W$?


  • $U = \langle v _ 1, \ldots, v _ m, u _ 1, \ldots, u _ p\rangle$
  • $W = \langle v _ 1, \ldots, v _ m, w _ 1, \ldots, w _ q\rangle$

The dimension formula states:

Let $U, W$ be subspaces of a finite dimensional vector space $V$. Then $\dim(U+W) + \dim(U \cap W) = \dim(U) + \dim(W)$.

Given that

  • $U \cap W = \langle v _ 1, \ldots, v _ m \rangle$
  • $U = \langle v _ 1, \ldots, v _ m, u _ 1, \ldots, u _ p\rangle$
  • $W = \langle v _ 1, \ldots, v _ m, w _ 1, \ldots, w _ q\rangle$

What do you need to show is a valid basis for $U + W$ in order to prove the result?


\[\\{v_1, \ldots, v_m,u_1,\ldots,u_p,w_1,\ldots,w_q\\}\]

What two things do you need to show for a set to be a basis of a vector space?


  • Spanning
  • Linearly independent

The difficult part of proving

\[\dim(U+W) +\dim(U\cap W) = \dim(U) + \dim(W)\]

comes down to proving

\[\\{v_1, \ldots, v_m,u_1,\ldots,u_p,v_1,\ldots,v_q\\}\]

is linearly independent (and spanning, and hence a basis). What do you have to notice once you’ve rearranged

\[\sum^m_{i=1}\alpha_i v_i + \sum_{i=1}^{p}\beta_i u_i = -\sum^q_{i=1}\gamma_iw_i\]

That this vector must in $U \cap W$.




Related posts