A discriminant is a scalar invariant associated to a polynomial, algebraic form, or algebraic object that vanishes precisely when the object has singularities such as repeated roots, multiple factors, or non-transverse intersections.

Principle

It encodes degeneracy: for univariate polynomials it equals (up to scale) the product of squared differences of roots and can be expressed as the determinant of a Sylvester or discriminant matrix; it behaves predictably under scaling and field extension.

Demonstration

For a quadratic ax^2 + bx + c the discriminant is Δ = b^2 − 4ac, which vanishes exactly when the polynomial has a double root; for a cubic it is a degree-4 polynomial in the coefficients that vanishes on multiple-root loci.

Misapplication

Confusing polynomial discriminant with field discriminant or using the univariate discriminant formula naively in multivariate or geometric settings where the relevant discriminant is a different, often much more complicated, object.

Consequence

When computed correctly, the discriminant detects singular or ramified fibers, controls separability, and guides factorization and resolution of singularities; a nonzero discriminant often signals generic smoothness.

Reversal

The complement perspective treats nonvanishing discriminant as generic separability or transversality: reversing the condition isolates the locus of regular, distinct-root or non-singular objects.

Boundary

Definition and form of discriminant depend on the object: univariate polynomial discriminants differ from discriminants of forms, number fields, or multivariate systems; over characteristic p there are subtleties with separability and vanishing behaviors.

Semantic Tension

Tension exists between different 'discriminants' (polynomial vs field vs form discriminant): they share the idea of detecting degeneracy but differ in algebraic construction, scaling, and geometric meaning, which can cause confusion.

Synthesis

The discriminant is the algebraic detector of degeneracy for an object—vanishing precisely on singular or multiply-rooted cases—and must be interpreted in the correct category (polynomial, form, field, or system) to yield meaningful conclusions.