TheoremComplete

Classification Theorems

Classification theorems for bilinear and quadratic forms provide complete invariants determining when two forms are equivalent. The answer depends crucially on the underlying field.

TheoremClassification over $\mathbb{R}$

Two real symmetric bilinear forms (or quadratic forms) are congruent if and only if they have the same signature (p,q,r)(p, q, r).

This means every real quadratic form is congruent to exactly one of: i=1pyi2i=p+1p+qyi2\sum_{i=1}^p y_i^2 - \sum_{i=p+1}^{p+q} y_i^2

The rank is p+qp + q, and signature completely classifies the form.

TheoremClassification over $\mathbb{C}$

Over complex numbers, symmetric bilinear forms are classified only by rank. Every non-degenerate form of rank rr is congruent to: y12+y22++yr2y_1^2 + y_2^2 + \cdots + y_r^2

The distinction between positive/negative disappears since 1=i2-1 = i^2 is a square in C\mathbb{C}.

TheoremWitt's Cancellation Theorem

If V1WV2WV_1 \oplus W \cong V_2 \oplus W as spaces with bilinear forms (where WW is non-degenerate), then V1V2V_1 \cong V_2.

This allows "canceling" common summands when comparing forms, analogous to cancellation in arithmetic.

ExampleConic Section Classification

In R2\mathbb{R}^2, quadratic forms Q(x)=ax12+2bx1x2+cx22Q(\mathbf{x}) = ax_1^2 + 2bx_1x_2 + cx_2^2 classify conics via discriminant Δ=b2ac\Delta = b^2 - ac:

  • Δ<0\Delta < 0: Ellipse (signature (2,0)(2,0) or (0,2)(0,2))
  • Δ>0\Delta > 0: Hyperbola (signature (1,1)(1,1))
  • Δ=0\Delta = 0: Parabola (degenerate, rank 1)

The signature provides geometric classification.

TheoremSymplectic Forms

A skew-symmetric bilinear form ω\omega is called symplectic if non-degenerate. Every symplectic form on R2n\mathbb{R}^{2n} is congruent to: ω0=i=1n(xidyiyidxi)\omega_0 = \sum_{i=1}^n (x_idy_i - y_idx_i)

This canonical form underlies Hamiltonian mechanics and symplectic geometry. Dimension must be even; symplectic forms exist only on even-dimensional spaces.

Remark

Classification theorems transform infinite families of matrices into finite lists of canonical forms. Over R\mathbb{R}, geometry matters: signature distinguishes elliptic from hyperbolic behavior. Over C\mathbb{C}, only rank survives. These results extend to more general settings (p-adic fields, function fields) with increasingly subtle invariants, forming the foundation of algebraic K-theory and quadratic form theory.