Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $V_1, \ldots, V_n$ be $n$ subspaces of a vector space $V$.

Is there a formula for $\dim(V_1 + \cdots + V_n)$ similar to $\dim(V_1 + V_2)=\dim(V_1) + \dim(V_1) - \dim(V_1 \cap V_2)$?

share|cite|improve this question
Didn't notice those, thanks for referring. – Peter P. Jan 26 '12 at 13:24
As a curiosity: At the moment the answer related to this is the most upvoted in the MO thread Examples of common false beliefs in mathematics – Martin Sleziak Jan 26 '12 at 16:09
up vote 6 down vote accepted

This example will illustrate the difficulty.

Let $i,j,k$ be the standard basis vectors for 3-dimensional real space. Let $W,X,Y,Z$ be the subspaces spanned by $i,j,k,i+j$, respectively. Then $W+X+Y$ has dimension 3, $W+X+Z$ has dimension 2, but $W,X,Y,Z$ all have dimension 1, and any intersection of two or more has dimension 0. So, no formula using dimensions of the four spaces and their intersections can distinguish between $W+X+Y$ and $W+X+Z$.

share|cite|improve this answer

You can do it iteratively using $$\dim(V_1 + V_2) = \dim(V_1) + \dim(V_2) - \dim(V_1 \cap V_2)$$ i.e., set $V_1' = V_1$, then $V_2' = V_1' + V_2$, so you get $$\dim(V_2') = \dim(V_1') + \dim(V_2) - \dim(V_1' \cap V_2)$$ and so on, let $V_i' = V_{i-1}' + V_i$ so $$\dim(V_i') = \dim(V_{i-1}') + \dim(V_i) - \dim(V_{i-1}' \cap V_i)$$

If you want to, you can substitute the formula for $\dim (V_{i-1}')$ in the method above to get a formula, e.g. if you have $V_1, V_2, V_3$: $$\dim(V_2') = \dim(V_1) + \dim(V_2) - \dim(V_1 \cap V_2)$$ $$\begin{align*} \dim(V_3') &= \dim(V_2') + \dim(V_3) - \dim(V_2' \cap V_3) = \\ &=\dim(V_1) + \dim(V_2) - \dim(V_1 \cap V_2) + \dim(V_3) - \dim( (V_1 + V_2) \cap V_3) \end{align*}$$ and of course $V_3' = V_1 + V_2 + V_3$, so this yields a formula for the dimension. The same method can be applied to an arbitrary finite number of subspaces, which yields $$\begin{align*} \dim \left( \sum_{i = 1}^n V_i \right) &= \dim \left( \sum_{i = 1}^{n-1} V_i \right) + \dim(V_n) - \dim \left( \sum_{i = 1}^{n-1} V_i \cap V_n \right) = \\ &= \dots = \sum_{i = 1}^n \dim(V_i) - \sum_{j = 1}^{n-1} \dim \left( \sum_{k = 1}^j V_k \cap V_{j+1} \right) \end{align*}$$

share|cite|improve this answer
In the last line, surely you mean finite number of subspaces? – Ravi Jan 26 '12 at 15:22
Yeah, sure. Corrected. – Calle Jan 26 '12 at 15:25
In the last line, why do we have $- \dim \left(\sum_{i=1}^{n-1} V_i \cap V_n \right)$ instead of $- \dim \left( V_n\cap\sum_{i=1}^{n-1} V_i\right)$? – Leon Feb 8 at 22:55
@Leon, they are the same thing. – Calle Feb 9 at 22:35

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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