A representation of a function or solution as a series (often formal) whose terms are ordered by magnitude in a limiting regime of a parameter, so that truncating the series after finitely many terms yields an approximation whose error is smaller than the last retained term in that regime.

Principle

Order terms by their size in the limit of a small or large parameter and retain leading contributions while controlling the remainder relative to the limiting parameter.

Demonstration

Stirling's asymptotic series: log Γ(n+1) ~ n log n − n + (1/2)log(2π n) + 1/(12n) − 1/(360n^3) + … gives increasingly accurate approximations for large n when truncated appropriately.

Misapplication

Treating an asymptotic series as convergent for all parameter values or summing divergent asymptotic terms as if they produced a uniformly valid convergent sum outside the intended limit regime.

Consequence

Provides controlled approximations that capture leading behaviour and quantify the size of neglected terms in the intended limit, often enabling analytic estimates and simplified computations.

Reversal

A convergent power series expansion valid in a finite radius of convergence (Taylor series) that guarantees term-by-term convergence near a point rather than only ordering by limit-size.

Boundary

Applies when a clear limiting regime (e.g., parameter →0 or →∞) exists; may be formal or divergent, and accuracy depends on truncation and the parameter's size; not a universal replacement for uniform convergence or exact solutions.

Semantic Tension

Tension with convergent series: asymptotic expansions prioritize correct limiting behavior and term ordering over convergence, whereas convergent series prioritize pointwise or uniform convergence properties.

Synthesis

An asymptotic expansion is a limit-focused series representation that orders contributions by magnitude to yield practical, often nonconvergent, approximations whose truncation error is controlled relative to the small or large parameter.