Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

I have read on this site and from various sources trying to get an understanding of what the change of basis matrix and change of basis operator is. Below, I have given the definitions as I understand them and hope that someone can either confirm or correct my understanding of these things.

Part I

Let $v := (v_1, \dots, v_n)$ and $w := (w_1, \dots, w_n)$ be two bases of a vector space $V$. By a theorem, there exists a unique isomorphism $T:V \rightarrow V$ given by $T(v_i) = w_i, i=1, \dots n$. Let $[T]^v_w$ denote the matrix of $T$ with respect to $v$ and $w$. Then for any $x \in V$, by another theorem, $$ [T]^v_w[x]_v = [x]_w $$ where $[x]_b$ denotes the column matrix of $x$ with respect to a basis $b$. The matrix $[T]^v_w$ is called the change of basis matrix or the transition matrix from $v$ to $w$ and $T$ is called the transition automorphism or transition operator from $v$ to $w$.

Question: Is everything in the above acceptably defined?

Part II Looking at the above componentwise, for a vector $x = a^1v_1 + \dots +a^nv_n$ expressed in terms of the basis $v$, the $k$-th coordinate of $[x]_w$ is given by $$ ([x]_w)_k = \sum_{j = 1}^n T^k_ja^j $$ where $T^k_j$ denotes the entry of $T$ that lives in the $k$-th row and $j$-th column of $[T]^v_w$.

Question: How are the actual basis vectors related to one another with respect to the components of $T$? I think it should be $w_k = \sum_{k=1}^n T^j_k v_j$. How does one prove this rigorously? I believe it follows from the fact that the $T(v_k)$ corresponds to the $k^th$ column of $T$ which is represented by the scalars $T^j_k, k=1, \dots, n$. Is this right?

share|cite|improve this question
Your last question: No; what you have is that $v_k = \sum T_k^jw_j$. To see this, note simply that $[v_k]_v$ has a $1$ in the $k$th entry and $0$s elsewhere; this will give you $[v_k]_w$. To express the $w_i$ in terms of the $v_j$'s, you need $T^{-1}$. – Arturo Magidin Jul 3 '12 at 18:05
@ArturoMagidin So, $w_k = \sum^n_{k=1} (T^{-1})^j_k v_j$? – AFX Jul 3 '12 at 18:08
Yes; just switch the roles of $v_i$ and $w_j$ when using $T^{-1}$. – Arturo Magidin Jul 3 '12 at 18:08
Side note: in future, it's much easier if you split a long question into several questions posted over spaced time period. It works better & you'll get more (quantity & quality) feedback. – user2468 Jul 3 '12 at 18:33
@ArturoMagidin Thanks for the tips but I just don't see it; it seems backward to me. – AFX Jul 3 '12 at 18:59

The elements of $j$-th column of $T$, $t_{1j}, t_{2j}, \cdots, t_{nj}$, are the expansion coefficients of the $j$-th basis $v_j$ with respect to new basis $w_1,\cdots, w_n$, i.e. $v_j = \sum_{i=1}^nt_{i,j}w_i$. Likewise, the elements of $i$-th row of $T$, $t_{i1}, t_{i2}, \cdots, t_{in}$, are the expansion coefficients of the $i$-th basis $w_i$ with respect to old basis $v_1,\cdots, v_n$, i.e. $w_i = \sum_{j=1}^nt_{i,j}v_j$.

share|cite|improve this answer

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.