Determinant — Physica

Equivalent forms

\det(AB) = \det A \,\det B

\det(A) = \prod_i \lambda_i

A single scalar that simultaneously answers 'invertible?', 'how much does volume scale?', and 'is orientation flipped?'.

Symbol	Name	SI unit	Dimension	Range
$a$	Entry a Top-left	1	1	-2 – 2
$b$	Entry b Top-right (shear)	1	1	-2 – 2
$c$	Entry c Bottom-left (shear)	1	1	-2 – 2
$d$	Entry d Bottom-right	1	1	-2 – 2
$\det$	Determinantoutput Signed area scale factor	1	1	-8 – 8

Unit systems

Symbol	SI	CGS	Imperial
$a$	1	1	1

Where it holds

Defined for any square matrix; the geometric area/volume picture holds over the reals. Generalizes to

n\times n

via cofactor or permutation sum.

Dimensional analysis

det (pure number here)

ad - bc

With dimensionless entries the determinant is dimensionless; as a Jacobian it carries the product of the coordinate dimensions.

\checkmark

Discovery

Leibniz, Cauchy · 1812

Leibniz used determinants for solving linear systems in 1693; Seki in Japan independently earlier. Cauchy's 1812 memoir systematized them and proved det(AB)=det(A)det(B). They became the gateway to eigenvalues, Jacobians, and orientation.

Try this

One number tells you whether a matrix stretches space, squashes it flat, or flips it inside out — and whether the equations it encodes can be solved at all. That number is the determinant.

Deform the unit square with a 2×2 matrix and read off the area of the resulting parallelogram; when it collapses to zero the matrix is singular.

Research status: stable

Real-world applications

Testing solvability/invertibility of linear systems
Jacobian factors in change-of-variables integrals
Orientation and chirality in geometry
Volume of parallelepipeds and cross products

Common misconceptions

det measures size of the matrix — it measures volume scaling of the map.
A big determinant means a 'stable' matrix — conditioning, not determinant, governs numerical stability.
det(A+ $B) = det$ A + det B — false; determinants are multiplicative, not additive.

Experimental verification

Change-of-variable integrals in physics (e.g. spherical coordinates) carry the Jacobian determinant; matching computed and measured volumes/fluxes confirms it. Orientation flips predict parity behavior seen in optics and crystallography.

Derivation

The columns of A are the images of the basis vectors.

The parallelogram they span has signed area ad - bc (base

\times

height with orientation), which is exactly the determinant.

For

n\times n

it is the signed n-volume of the image of the unit cube.

Limiting cases

det = 0

⟶ singular matrixColumns are linearly dependent; the map flattens the plane onto a line.

det < 0⟶ orientation flippedThe transformation includes a reflection.

det = 1

⟶ area-preservingSpecial linear group SL(2) — shape may distort but area is fixed.

What if…

What if I swap two rows?

The determinant flips sign — a reflection of orientation.

What if I scale one row by k?

The determinant scales by k; scaling the whole $n\times n$ matrix scales det by k^n.

1

Determinant of a shear

Given ·

A:: [[1,2],[0,1]]

Find · det

Steps

$det = (1)(1) - (2)(0)$
$= 1 - 0$
$= 1$ : shear preserves area

Result · 1

2

Singular check

Given ·

A:: [[2,4],[1,2]]

Find · invertible?

Steps

$det = 2\cdot 2 - 4\cdot 1$
$= 4 - 4 = 0$
Rows are proportional $\to not$ invertible

Result · no

(det = 0)

Eigenvalue Equation→matrix inversejacobian determinantRotation Matrices→