Skip to content

Legendre Polynomials

The uniquely normalized polynomial sequence orthogonal on [-1,1] with unit weight and solving Legendre's differential equation.

Version
v1 · 2026-08-30 · History
Domain-specific #
2174
Origin domain
mathematics
Subdomain
special functions
Aliases
Legendre polynomial sequence

Core Idea

The Legendre polynomials \(P_n(x)\) are the normalized sequence of real polynomials, one of each nonnegative degree, orthogonal on \([-1,1]\) with respect to unit weight. Their standard normalization is \(P_n(1)=1\). Equivalently, \(P_n\) is the polynomial solution of Legendre’s differential equation

\[ (1-x^2)y''-2xy'+n(n+1)y=0 \]

with that normalization. The two characterizations connect spectral expansion, boundary-value problems, spherical harmonics, and Gaussian quadrature.[1][2]

Orthogonality is exact:

\[ \int_{-1}^{1}P_m(x)P_n(x)\,dx=\frac{2}{2n+1}\delta_{mn}. \]

Rodrigues’ formula, the three-term recurrence, and the generating function are equivalent access routes to the same sequence. The abstraction is not “any orthogonal polynomial.” Its identity fixes interval, weight, normalization, differential operator, and recurrence together.

Structural Signature

Mandatory roles:

  • The degree index \(n\ge0\) selects a polynomial of exact degree \(n\).
  • The interval \([-1,1]\) is the standard real orthogonality domain.
  • The unit weight defines the inner product \(\langle f,g\rangle=\int_{-1}^1fg\,dx\).
  • The normalization \(P_n(1)=1\) fixes scale.
  • Legendre’s differential operator has eigenvalue \(n(n+1)\).
  • The three-term recurrence connects adjacent degrees.
  • Completeness in the relevant square-integrable space licenses Legendre-series expansion.

Recognition test. A polynomial family qualifies when these roles agree under a stated equivalent normalization. Shifted Legendre polynomials on \([0,1]\), associated Legendre functions, and orthonormal rescalings are variants or relatives whose transformations must be declared.

What It Is Not

  • It is not the Legendre transform from convex analysis; the shared name does not imply a functional transform.
  • It is not every polynomial solution of the differential equation without degree and normalization conditions.
  • It is not the associated Legendre family \(P_n^m\), which adds an order and generally is not polynomial in the same sense for arbitrary conventions.
  • It is not a spherical harmonic, although the zonal angular factor uses \(P_n(\cos\theta)\).
  • It is not a generic polynomial basis. Monomials and Chebyshev polynomials have different conditioning, weights, and operators.

Scope of Application

Legendre polynomials recur wherever functions on a finite interval are expanded under uniform weight. In numerical analysis they form spectral bases and determine Gauss–Legendre quadrature nodes through their zeros. In potential theory the generating function expands the Newtonian kernel in separated radial and angular variables. In mathematical physics they solve the azimuthally symmetric angular part of Laplace’s equation in spherical coordinates. In approximation, truncated Legendre series project data onto mutually orthogonal degrees.

The standard sequence can be rescaled to another finite interval, but the affine map and normalization must be carried. A problem with endpoint-singular weights naturally calls for Jacobi or Chebyshev polynomials instead; calling it Legendre merely because the domain is finite loses the defining weight.

The same sequence also mediates between continuous analysis and discrete computation. An analyst may define a projection by integration, a numerical code may generate values by recurrence, and a quadrature routine may work from zeros and weights. These are not separate abstractions merely sharing a name: their compatibility is secured by the common normalization and differential-equation identity. Conversely, a modified recurrence that loses exact orthogonality should be described as an approximation scheme rather than silently redefining the family.

Clarity

The named family aligns several formulas that otherwise appear unrelated. Rodrigues’ formula

\[ P_n(x)=\frac{1}{2^n n!}\frac{d^n}{dx^n}(x^2-1)^n \]

fixes polynomial coefficients and parity. The recurrence

\[ (n+1)P_{n+1}(x)=(2n+1)xP_n(x)-nP_{n-1}(x) \]

supports stable sequential generation. Orthogonality extracts coefficients independently. Agreement among these representations is a powerful implementation check; disagreement usually reveals a normalization or indexing error.

Manages Complexity

Orthogonality converts a coupled approximation problem into scalar projections. Given a suitable function \(f\), its coefficient is

\[ a_n=\frac{2n+1}{2}\int_{-1}^{1}f(x)P_n(x)\,dx. \]

A truncated expansion isolates approximation by degree and avoids solving a dense normal equation in exact arithmetic. The recurrence avoids repeatedly expanding Rodrigues derivatives. Gauss–Legendre quadrature reuses zeros and weights to integrate polynomials of high degree exactly with relatively few samples.

For repeated transforms, precomputed nodes, weights, and recurrence coefficients replace symbolic manipulation. Symmetry halves some work because odd and even degrees behave predictably under reflection. These efficiencies depend on the exact family: importing weights from a Chebyshev rule or using a shifted interval without converting coordinates destroys the advertised exactness.

This compression has limits. Truncation error depends on regularity, and finite-precision orthogonality can degrade at high degree. The family manages functional variation, not boundary singularities or arbitrary geometries by itself.

Abstract Reasoning

Parity follows from Rodrigues’ formula: \(P_n(-x)=(-1)^nP_n(x)\). Orthogonality follows from integration by parts because lower-degree polynomials vanish after enough derivatives while endpoint terms vanish. The differential equation makes degrees distinct eigenfunctions of a Sturm–Liouville operator, explaining orthogonality structurally rather than as a table of integrals.

Zeros of \(P_n\) lie inside \((-1,1)\), are simple, and interlace those of adjacent degrees. These properties support quadrature and root-finding diagnostics. A claimed \(P_n\) with wrong parity, incorrect value at one, or a zero outside the interval fails the standard identity.

The generating function \((1-2xt+t^2)^{-1/2}=\sum_{n\ge0}P_n(x)t^n\), for its convergence range, provides another consistency route. Differentiating or comparing coefficients recovers recurrence relations and connects the family to multipole expansions. The point is not to accumulate formulas; it is that independently derived representations constrain the same object.

Knowledge Transfer

Transfer is literal across approximation, quadrature, and spherical boundary problems because the same sequence and invariants are used. Coefficients computed in one Legendre basis can be interpreted in another application after matching the interval and normalization.

The portable skeleton is Orthogonal Basis: choose an inner product, construct mutually orthogonal elements, and project. That skeleton transfers to Fourier, Chebyshev, Hermite, and wavelet systems. The Legendre name should not transfer unless unit weight on the mapped interval and the corresponding polynomial family remain present.

Examples

Low degrees. The recurrence gives \(P_0(x)=1\), \(P_1(x)=x\), \(P_2(x)=(3x^2-1)/2\), and \(P_3(x)=(5x^3-3x)/2\). Each satisfies \(P_n(1)=1\), the required parity, and orthogonality to all lower-degree polynomials under unit weight.[3]

Projection. For \(f(x)=x^2\), symmetry eliminates odd coefficients. Since \(x^2=(1+2P_2(x))/3\), only degrees zero and two are needed. This exact decomposition exhibits basis role, coefficient extraction, and degree separation.

Quadrature. The two roots of \(P_2\), \(\pm1/\sqrt3\), are the nodes of two-point Gauss–Legendre quadrature, which integrates every polynomial through degree three exactly. The quadrature rule is an application, not the polynomial family’s entire identity.

Structural Tensions

  • Canonical normalization versus orthonormal convenience: \(P_n(1)=1\) is standard, while numerical work often rescales to unit norm. Diagnostic: is the normalization stated before coefficients are compared?
  • Exact orthogonality versus finite precision: mathematical inner products vanish, but computed recurrences accumulate error. Diagnostic: does the numerical Gram matrix remain acceptably close to diagonal?
  • Global efficiency versus local irregularity: smooth functions converge rapidly, while discontinuities create slow coefficient decay and oscillation. Diagnostic: do tail coefficients decrease at the predicted rate?
  • Equivalent definitions versus convention drift: differential equation, recurrence, and Rodrigues formula agree only with consistent signs and scales. Diagnostic: do all definitions yield \(P_0=1\), \(P_1=x\), and \(P_n(1)=1\)?
  • Autonomy versus generic basis: Orthogonal Basis supplies the skeleton but not the unit weight or Legendre operator. Diagnostic: after removing interval, weight, and operator, does any Legendre-specific content remain?

Structural–Framed Character

Legendre Polynomials are strongly structural. Their identity is fixed by equations, normalization, and an inner product. Recognition does not depend on institution or historical interpretation. Applications supply frames, but the same polynomial objects survive unchanged across them.

They remain domain-specific rather than prime because their literal structure is a specialized mathematical family. The cross-domain pattern is Basis or Orthogonality, not the particular \(P_n\) sequence.

Structural Core vs. Domain Accent

Structural core. Mutually orthogonal indexed elements support projection, recurrence, eigenfunction reasoning, and approximation.

Domain accent. Polynomial degree, \([-1,1]\), unit weight, Legendre’s differential equation, Rodrigues’ formula, and the normalization at one uniquely identify this family. Removing those features leaves a general orthogonal-basis abstraction.

The residual is substantial enough for an autonomous domain-specific node because it supplies formulas, invariants, and computational consequences that Basis alone does not entail.

Legendre Polynomials instantiate Basis: finite initial segments span polynomials up to the matching degree, and the completed orthogonal system supports suitable function expansions. They relate to Orthogonality, Recursion, and Eigenvalue and Eigenvector. Those abstractions explain portions of the structure, but none entails the exact sequence.

Relationships to Other Abstractions

Local relationship map for Legendre PolynomialsParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Legendre PolynomialsDOMAINPrime abstraction: Basis — presupposesBasisPRIME

Current abstraction Legendre Polynomials Domain-specific

Parents (1) — more general patterns this builds on

  • Legendre Polynomials presupposes Basis Prime

    Legendre Polynomials instantiate Basis: finite initial segments span polynomials up to the matching degree, and the completed orthogonal system supports suitable function expansions.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Legendre Polynomials sits in a sparse region of the domain-specific corpus (86th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Chebyshev polynomials: orthogonal under a singular endpoint weight. Tell: is the weight exactly one?
  • Associated Legendre functions: carry an additional order \(m\). Tell: does the angular equation include the \(m^2/(1-x^2)\) term?
  • Shifted Legendre polynomials: live on \([0,1]\). Tell: has the affine variable map been applied?
  • Legendre transformation: maps a function to a conjugate-variable representation. Tell: is the object an indexed polynomial sequence or a convex dual operation?
  • Spherical harmonics: combine associated functions with azimuthal factors. Tell: is a two-angle basis being used?

References

[1] NIST Digital Library of Mathematical Functions, §18.3, “Definitions: Classical Orthogonal Polynomials,” https://dlmf.nist.gov/18.3. registry ↩

[2] NIST Digital Library of Mathematical Functions, §18.9, “Recurrence Relations and Derivatives,” https://dlmf.nist.gov/18.9. registry ↩

[3] NIST Digital Library of Mathematical Functions, §18.5, “Explicit Representations,” https://dlmf.nist.gov/18.5. registry ↩