Linear Algebra 32 | Transposition for Matrices

TL;DR

Transposition swaps rows and columns, rearranging entries without changing their values.

Briefing Cornell Notes

Briefing

Matrix transposition is the operation that swaps a matrix’s rows and columns, turning an M×N matrix into an N×M matrix while preserving the same entries—just rearranged. For vectors, the same idea appears as turning a column vector into a row vector (and back again). This matters because many calculations, especially those involving inner products and matrix products, depend on how row/column orientation affects multiplication.

For a column vector a, transposition is denoted by a^T and converts it into a row vector. Applying transposition twice returns the original object, so (a^T)^T = a. The transcript extends this notation to matrices: if A is an M×N matrix, then A^T is defined to be an N×M matrix. The (i,j)-entry of A^T comes from the (j,i)-entry of A, which is why transposition can be pictured as reflecting the matrix across its main diagonal. Under this reflection, the j-th column of A becomes the j-th row of A^T, and the j-th row of A becomes the j-th column of A^T.

Concrete examples make the shape change clear. A 2×2 matrix transposes into another 2×2 matrix, so the matrix becomes “taller” or “wider” only when the original is rectangular. For a square matrix, the diagonal reflection stays within the same dimensions. The transcript gives an explicit 2×2 example where A^T = [[1,3],[2,4]] when A is arranged so that the off-diagonal entries swap positions.

A special case occurs when transposition leaves the matrix unchanged: if A^T = A, the matrix is symmetric. Symmetric matrices are exactly those whose entries mirror across the main diagonal, meaning the matrix looks the same after the reflection. The discussion highlights that this property is tied directly to the diagonal symmetry of the array of numbers.

Finally, transposition interacts with matrix multiplication in a way that affects how expressions must be rearranged. If A and B can be multiplied (so AB is defined), then the transpose of the product reverses the order of factors: (AB)^T = B^T A^T. This reversal follows from the row/column swap inherent in transposition and is presented as an important rule to remember. The transcript closes by pointing toward later material connecting transposition to inner products and symmetric matrices.

Cornell Notes

Transposition swaps rows and columns. For a vector a, the transpose a^T turns a column vector into a row vector, and doing it twice returns the original: (a^T)^T = a. For a matrix A with size M×N, the transpose A^T has size N×M, with entries flipped across the main diagonal: the j-th column of A becomes the j-th row of A^T (equivalently, (A^T)_{ij} = A_{ji}). Square matrices transpose into the same shape, and if A^T = A, the matrix is symmetric. A key multiplication rule is (AB)^T = B^T A^T, which reverses the order of factors because transposition swaps row/column roles.

What does transposition do to a column vector, and why does applying it twice return the original?

Transposition takes a column vector a and produces a row vector a^T. The entries stay the same, but their orientation changes from “standing up” to “lying across.” Because transposition simply swaps the roles of rows and columns, doing it again swaps them back, giving (a^T)^T = a.

How is the transpose of an M×N matrix defined, and what happens to its dimensions?

If A is an M×N matrix, then A^T is defined to be an N×M matrix. The entries are rearranged by reflecting across the main diagonal: the j-th column of A becomes the j-th row of A^T, and the j-th row of A becomes the j-th column of A^T.

How can you compute A^T from A without rewriting everything from scratch?

Use the diagonal reflection rule. For each position (i,j) in A, the entry moves to position (j,i) in A^T. In other words, the (i,j)-entry of A^T is the (j,i)-entry of A.

Why are transposes of square matrices especially straightforward?

Square matrices have the same number of rows and columns (M=N), so swapping rows and columns keeps the matrix’s shape unchanged. The transpose still reflects across the main diagonal, but the dimensions remain identical, making it easy to visualize and compute.

What makes a matrix symmetric, and how does transposition characterize it?

A matrix is symmetric exactly when transposition does not change it: A^T = A. That means the matrix’s entries match their mirror images across the main diagonal, so the reflection leaves the array unchanged.

What is the rule for transposing a matrix product AB?

If AB is defined, then (AB)^T = B^T A^T. Transposition swaps row/column roles, which forces the order of multiplication to reverse when taking the transpose of the product.

Review Questions

Given a 3×2 matrix A, what are the dimensions of A^T, and where does the entry A_{2,1} go in A^T?
If A is symmetric, what equality must hold between A and A^T?
For compatible matrices A and B, write the formula for (AB)^T and explain what happens to the order of factors.

Key Points

1
Transposition swaps rows and columns, rearranging entries without changing their values.
2
For vectors, a^T converts a column vector into a row vector, and (a^T)^T = a.
3
If A is M×N, then A^T is N×M.
4
The transpose is a diagonal reflection: (A^T)_{ij} = A_{ji}.
5
Square matrices keep the same shape after transposition, making computation simpler.
6
A matrix is symmetric exactly when A^T = A.
7
Transposing a product reverses factor order: (AB)^T = B^T A^T.

Highlights

Transposition turns an M×N matrix into an N×M matrix by reflecting across the main diagonal.

For vectors, transposition flips orientation (column to row) and twice returns the original: (a^T)^T = a.

Symmetric matrices are precisely those unchanged by transposition: A^T = A.

The transpose of a product reverses multiplication order: (AB)^T = B^T A^T.

Topics

Matrix Transpose
Row/Column Swap
Symmetric Matrices
Matrix Product Transpose
Diagonal Reflection