A family of arithmetic measures that assign a nonnegative real number to algebraic numbers or rational points on varieties, quantifying their arithmetic complexity and growth. Common examples include naive heights on projective coordinates, the Weil height, and canonical heights such as the Néron–Tate height on abelian varieties.

Principle

Track and compare arithmetic size by aggregating contributions from all places of a number field, using logarithmic valuations or homogeneous coordinate norms to obtain height functions that are functorial up to bounded error.

Demonstration

For a rational point on projective space represented by coprime integers (x0:...:xn), the naive projective height is log max(|x0|,...,|xn|). On an elliptic curve, the Néron–Tate canonical height is a quadratic form that grows like the square of multiplication and vanishes precisely on torsion points.

Misapplication

Using a height defined only at archimedean places to bound global Diophantine solutions, thereby ignoring p-adic contributions that dominate in some families and producing incorrect finiteness claims.

Consequence

When used correctly, heights enable explicit finiteness results, Northcott-type compactness statements, effective descent arguments, and quantitative bounds on solutions to Diophantine equations.

Reversal

Ignoring heights yields purely qualitative or local statements about points without any measure of arithmetic size; one then cannot control distribution, prove finiteness, or compare families effectively.

Boundary

Heights measure arithmetic complexity, not geometric complexity; they require choice of embedding or line bundle and are defined up to O(1) changes under different models. They do not automatically provide algorithmic bounds without auxiliary estimates.

Semantic Tension

Heights compete with naive size notions and with purely local valuations: naive coordinate size is simpler but not intrinsic, while local valuations give fine structure but must be assembled to a global height.

Synthesis

Height functions are global arithmetic size measures assembled from local valuations or norms; they turn qualitative arithmetic properties into quantitative inequalities that underpin finiteness, descent, and counting arguments.