The determinant is how much the matrix scales area or volume. For [[1,2],[3,4]] it is ad − bc = −2: the transformation doubles area and flips orientation. A determinant of zero means space gets squashed flat, and that is exactly when no inverse exists.
What it measures
Take the unit square. Apply the matrix. The determinant is the area of what comes out.
In three dimensions it is the volume of the transformed unit cube, and the same idea continues upwards. One number, and it says how much the transformation stretches or shrinks everything.
The sign carries the other half of the answer. A negative determinant means the transformation turned space inside out — a reflection somewhere in it. In two dimensions, anticlockwise becomes clockwise.
Reading it geometrically makes the rules stop being arbitrary. Why does swapping two rows flip the sign? Because it reflects the space. Why does scaling a row by k multiply the determinant by k? Because it stretches one direction by k.
What zero means
A determinant of zero is the most informative answer the calculation gives.
It means the transformation collapses space: a square becomes a line segment, a cube becomes a flat sheet. Volume goes to nothing.
And collapsing cannot be undone. Once a whole direction has been flattened to a point, no transformation recovers what was there — the information is gone rather than merely hard to get at. That is why a zero determinant and “no inverse” are the same statement.
It is also the same statement as “the rows are linearly dependent”: one row is a combination of the others, so they do not span enough directions to fill the space.
This is where exact arithmetic earns its place. In floating point, a determinant that ought to be zero comes out as 3e−17, and a page checking against zero declares the matrix invertible. Here the answer is exactly zero, and the singularity is a fact rather than a judgement call.
Why not cofactor expansion
Cofactor expansion — break the determinant into smaller ones along a row, with alternating signs — is how the topic is taught, and it is the right way to understand what a determinant is. It is a poor way to compute one.
It is O(n!). A 3×3 takes a handful of multiplications. A 5×5 takes about 120. An 8×8 takes around forty thousand, and a 20×20 would need more operations than there are seconds in the age of the universe.
Elimination is O(n³). The same 20×20 is a few thousand operations. That is not a constant-factor improvement; it is the difference between possible and impossible.
This page uses the fraction-free version of elimination, which has one further advantage on exact arithmetic: ordinary Gaussian elimination on an integer matrix produces fractions whose denominators grow at every step, while the fraction-free arrangement keeps every intermediate value a whole number. Each division it performs is exact by construction.
The properties worth knowing
A handful of facts turn most determinant questions into one line.
det(AB) = det(A) × det(B). The determinant of a product is the product of the determinants — which is obvious geometrically, since composing two transformations multiplies their volume scalings.
det(Aᵀ) = det(A). Transposing changes nothing.
A triangular determinant is the product of the diagonal. Everything else contributes nothing, which is exactly why elimination works: it makes the matrix triangular first.
Scaling one row by k multiplies the determinant by k — so scaling the whole n×n matrix multiplies it by kⁿ, not by k. That one catches people out.
Two identical rows give zero. They are dependent by definition, so the volume is flat.
Where it gets used
Invertibility. The single test for whether a matrix can be undone, and therefore whether a linear system has one solution.
Areas and volumes. The area of a triangle from its coordinates is half a determinant, and the shoelace formula is a sum of them.
Cramer’s rule. Solving a system as a ratio of determinants — elegant, and impractical beyond about 3×3 for exactly the O(n!) reason above.
Calculus. The Jacobian determinant is the volume factor in a change of variables, which is why it appears whenever an integral is transformed into polar or spherical coordinates.
Sources and methodology
Determinants and the algorithms for them are standard; these are the references.
Method. Computed by Bareiss fraction-free elimination on exact rationals. That combination matters twice over: exactness means a determinant of zero is a fact rather than a rounded near-zero, so singularity is detected rather than guessed at; and the fraction-free arrangement keeps every intermediate value an integer on an integer matrix, avoiding the denominator blow-up that ordinary Gaussian elimination causes on rationals. The suite checks it against an independently written cofactor expansion on every matrix up to 4×4. That engine is verified on every change against 104 hand-written assertions, including that fraction-free elimination agrees with an independently implemented cofactor expansion on every test matrix, and that det(AB) equals det(A)·det(B) exactly. The count and the per-case breakdown are published on the formula verification page.
Read the guide
A non-zero determinant is exactly the condition for an inverse to exist — the Matrix Inverse Calculator computes it when it does.