How this instrument works
For a 2×2 matrix A = [[a, b], [c, d]], the adjoint — more precisely the classical adjugate — is [[d, −b], [−c, a]]: the diagonal entries trade places and the off-diagonal entries change sign. In general the adjugate of an n×n matrix is the transpose of its cofactor matrix, built one minor at a time; a 2×2 matrix is the smallest case where every minor is a single leftover entry, so the whole cofactor expansion collapses into this four-symbol shortcut.
The name is worth pausing on, because it covers two different things. In most linear-algebra courses 'adjoint' means exactly this adjugate, the matrix behind A⁻¹ = adj(A) ⁄ det(A) — divide the adjugate by the determinant and you have inverted A without a single row operation. In functional analysis and quantum mechanics, though, 'adjoint' instead means the conjugate transpose, a wholly different object; if a formula you're reading suddenly involves complex conjugates, that's the other adjoint, not this one.
The adjugate is defined even when A itself cannot be inverted. If a, b, c, and d make det(A) = ad − bc equal to zero, adj(A) still exists and still satisfies A · adj(A) = det(A) · I — which now reads A · adj(A) = 0, the zero matrix. Nothing breaks; the identity that usually hands you an inverse instead reports, correctly, that no inverse exists.
- Enter the top row of your matrix into a (row 1, col 1) and b (row 1, col 2).
- Enter the bottom row into c (row 2, col 1) and d (row 2, col 2).
- Read Adjoint: row 1, col 1 (= d) and Adjoint: row 1, col 2 (= −b) for the top row of the result.
- Read Adjoint: row 2, col 1 (= −c) and Adjoint: row 2, col 2 (= a) for the bottom row.
- Divide all four outputs by ad − bc yourself to get A⁻¹, if that determinant isn't zero.
Worked example — inverting [[1, 2], [3, 4]]
Take a = 1, b = 2, c = 3, d = 4, the matrix [[1, 2], [3, 4]]. Swap the diagonal: d and a trade places, so the (1,1) entry becomes 4 and the (2,2) entry becomes 1. Negate the off-diagonal: b becomes −2 and c becomes −3. The adjugate comes out as Adjoint: row 1, col 1 (= d) = 4, Adjoint: row 1, col 2 (= −b) = −2, Adjoint: row 2, col 1 (= −c) = −3, and Adjoint: row 2, col 2 (= a) = 1 — the matrix [[4, −2], [−3, 1]].
That adjugate is not decoration. The determinant here is ad − bc = 1×4 − 2×3 = −2, so A⁻¹ = adj(A) ⁄ det(A) = [[4, −2], [−3, 1]] ⁄ (−2) = [[−2, 1], [1.5, −0.5]]. Multiply the original matrix by this inverse and the product is the identity matrix exactly, with no rounding anywhere in sight.
Questions
What's the difference between an adjoint and an adjugate matrix?
For a real matrix they're the same thing when 'adjoint' is used the classical way: adj(A) is the transpose of the cofactor matrix, and for 2×2 it reduces to swapping the diagonal and negating the off-diagonal. 'Adjoint' has a second, unrelated meaning in functional analysis and quantum mechanics — the conjugate transpose — so many modern texts prefer 'adjugate' specifically to dodge that clash.
Why does swapping the diagonal and negating the off-diagonal work?
Because the adjugate is the transpose of the cofactor matrix, and in a 2×2 matrix every cofactor is a single leftover entry rather than a sub-determinant. Deleting row 1 and column 1 leaves just d, so that cofactor is d; deleting row 1 and column 2 leaves c with a sign flip from (−1)^(1+2), giving −c. Transposing the resulting cofactor matrix is what puts −b and −c on opposite corners.
How does the adjugate matrix relate to the matrix inverse?
A⁻¹ = adj(A) ⁄ det(A) whenever det(A) is not zero — dividing every entry of the adjugate by the determinant is the standard hand method for inverting a 2×2 (or larger) matrix. The identity behind it, A · adj(A) = det(A) · I, holds for every square matrix regardless of whether that matrix happens to be invertible.
What's the most common mistake when computing an adjugate by hand?
Negating the wrong pair. Students often flip the sign of a and d, the diagonal that should just swap places, instead of b and c, the off-diagonal that should be negated. A quick check: applying the adjugate rule to the identity matrix should return the identity unchanged, since swapping 1 with 1 and negating 0's leaves it exactly as it was.
What happens to the adjugate when the matrix has no inverse?
It still exists. If det(A) = ad − bc equals zero, adj(A) is computed the same way — swap, negate — but A · adj(A) now equals the zero matrix instead of a scaled identity, since det(A) itself is zero. The adjugate doesn't fail; it correctly signals that A⁻¹ cannot be built, by turning the usual identity into 0 = 0.
Does a diagonal or symmetric matrix have any special adjugate shortcut?
A diagonal matrix's adjugate is itself: for [[5, 0], [0, 5]], swapping two equal diagonal entries changes nothing and negating zero leaves zero. A symmetric matrix's adjugate stays symmetric too — [[2, 3], [3, 2]] adjugates to [[2, −3], [−3, 2]] — but the shared off-diagonal value still gets negated, so symmetry survives while the actual numbers don't stay put.