Matrices as a Number System

Matrices follow rules that echo ordinary numbers. The identity matrix [[1,0],[0,1]] acts on any 2 by 2 matrix the way 1 acts on numbers: multiply by it and nothing changes. That makes it the natural partner for an inverse. Just as 5 times 1/5 gives 1, a matrix M times its inverse M⁻¹ gives the identity matrix.

Whether that inverse exists depends on one number: the determinant. For M = [[a,b],[c,d]], the determinant is ad - bc, the product of the main diagonal minus the product of the off diagonal. If that number is zero, M has no inverse at all, the matrix equivalent of trying to divide by zero. If it isn't zero, the inverse is built by swapping the diagonal entries, negating the off-diagonal entries, and dividing the whole result by the determinant: M⁻¹ = (1/(ad-bc)) [[d,-b],[-c,a]].

This is why matrix division doesn't exist as its own operation. Numbers divide because every nonzero number has a reciprocal; matrices instead multiply by an inverse when one exists, and that inverse can fail to exist even when the matrix isn't the zero matrix. This connects directly to systems of equations: a 2 by 2 matrix can encode two linear equations, and a zero determinant means its rows are proportional, so the lines it represents are either identical (infinite solutions) or parallel (no solution) — never a single clean answer.