How this instrument works
Transposing a matrix reflects every entry across its main diagonal: the number sitting in row i, column j trades places with the number in row j, column i, so that Tᵢⱼ = Aⱼᵢ. Nothing is added, scaled, multiplied, or discarded — the four figures that go into a 2×2 matrix are exactly the four figures that come back out, only reassigned to new addresses. Of the five basic matrix moves — addition, multiplication by a scalar, matrix multiplication, the trace, and this one — transpose is the only pure rearrangement; the others compute new values from the old ones, while this one just relabels where each value lives.
Picture the four cells as a small grid and draw a line from the top-left corner to the bottom-right — that line is the main diagonal, and transposing is a mirror-flip across it. Any entry sitting exactly on that line, A₁₁ or A₂₂, maps to itself, because row i and column i name the same cell whenever i equals j; only the off-diagonal pair, A₁₂ and A₂₁, actually trades places. Flip twice and everything lands back where it started — (Aᵀ)ᵀ = A always, which is why transpose is its own inverse rather than a one-way street like scaling a matrix by a constant.
A matrix that already equals its own transpose is called symmetric — [[2,3],[3,2]] is the smallest interesting example, since its two off-diagonal entries already match, so the flip changes nothing. This sheet only handles 2×2 grids, but the same rule scales to any rectangular matrix: transposing an m-by-n matrix produces an n-by-m matrix, so a 2-by-3 matrix becomes 3-by-2 — rows and columns literally trade dimensions, not just positions, once the matrix stops being square.
- Enter the top row of your matrix into row1,col1 and row1,col2.
- Enter the bottom row into row2,col1 and row2,col2, completing the 2×2 grid.
- Read the flipped result in the four Transpose fields: Transpose: row1,col1, Transpose: row1,col2, Transpose: row2,col1, and Transpose: row2,col2.
- Check the diagonal first — Transpose: row1,col1 should always match row1,col1, and Transpose: row2,col2 should always match row2,col2.
- Swap only the off-diagonal pair by hand as a sanity check: row1,col2 should reappear in Transpose: row2,col1, and row2,col1 should reappear in Transpose: row1,col2.
Worked example — transposing [[1, 2], [3, 4]]
Enter row1,col1 = 1, row1,col2 = 2, row2,col1 = 3, and row2,col2 = 4, the matrix [[1,2],[3,4]]. Transposing swaps the off-diagonal pair and leaves the diagonal alone: Transpose: row1,col1 stays 1 and Transpose: row2,col2 stays 4, while Transpose: row1,col2 reads 3 — pulled from row2,col1 — and Transpose: row2,col1 reads 2 — pulled from row1,col2. The result is the matrix [[1,3],[2,4]].
The two diagonal entries, 1 and 4, never move for any square matrix, because row i and column i point at the very same cell when i equals j — there is nothing to swap. Contrast that with scalar multiplication, which would change all four numbers, or matrix multiplication, which would combine this matrix's rows against another matrix's columns entirely; transpose alone leaves every value untouched and only reassigns its address.
Questions
What does transposing a 2×2 matrix actually do?
It reflects every entry across the main diagonal: the entry in row i, column j moves to row j, column i, so Tᵢⱼ = Aⱼᵢ. For [[1,2],[3,4]] that means the diagonal pair (1 and 4) stays put while the off-diagonal pair (2 and 3) trades places, giving [[1,3],[2,4]] — no value is created, scaled, or lost, only relocated.
Why do the diagonal entries never change when you transpose?
Because row i and column i name the exact same cell whenever i equals j — there is nothing on the other side to swap with. In [[1,2],[3,4]], the entries 1 and 4 sit on that line and map to themselves, so Transpose: row1,col1 and Transpose: row2,col2 always echo the original row1,col1 and row2,col2 fields untouched.
How is transpose different from matrix multiplication or the trace?
Transpose only rearranges one matrix's own entries and produces another matrix the same size; the trace ignores three of the four entries and returns a single number (the diagonal sum, 1 + 4 = 5 for this example); matrix multiplication combines two separate matrices' rows against columns into brand-new values. All three read the same four inputs completely differently.
Does transposing twice return the original matrix?
Yes — transpose is its own inverse. Apply it once to [[1,2],[3,4]] to get [[1,3],[2,4]], then transpose that result and every entry flips back across the diagonal a second time, landing on [[1,2],[3,4]] again. In formula terms, (Aᵀ)ᵀ = A always holds, for a matrix of any size.
What is a symmetric matrix, and how does it relate to transpose?
A symmetric matrix is one that already equals its own transpose, meaning its off-diagonal entries mirror each other before you even flip anything — [[2,3],[3,2]] is the simplest 2×2 example, since swapping the two matching 3s changes nothing. Diagonal matrices such as [[5,0],[0,5]] are symmetric for the same reason: the off-diagonal zeros have nothing to disturb.
What is the most common mistake when transposing by hand?
Swapping the wrong pair — moving the diagonal entries into the off-diagonal spots instead of leaving them alone, or treating the flip like a ninety-degree rotation rather than a mirror across the diagonal. A quick check: if a diagonal entry moved, or a non-square matrix's dimensions did not swap (an m-by-n matrix becoming n-by-m), the transpose was done wrong.