Introduction
The term local symbol arises in the study of linear partial differential operators and in microlocal analysis. It refers to the leading order part of a differential operator when viewed locally, typically as a function on the cotangent bundle of a smooth manifold. While the global symbol of an operator captures its full behavior across the manifold, the local symbol focuses on the behavior in a neighbourhood of a point, providing essential information for questions about ellipticity, regularity, and propagation of singularities. The concept is central to the symbolic calculus developed by Lars Hörmander and others, and it underlies modern treatments of elliptic and hyperbolic partial differential equations.
History and Background
Early Developments
Linear partial differential operators were first studied in the 19th century by mathematicians such as Fourier and Laplace. The notion of a symbol for a constant coefficient operator - its Fourier multiplier - was well known, but a systematic theory for variable coefficient operators did not emerge until the work of Laurent Schwartz and others in the early 20th century. In 1943, Schwartz introduced the idea of a symbol as a distributional kernel, and by the 1950s the language of distribution theory and Fourier analysis had matured sufficiently to handle variable coefficient operators.
Hörmander's Symbol Calculus
Lars Hörmander made the concept of a local symbol precise in the 1960s, developing a robust pseudodifferential calculus that included symbol classes, quantization maps, and composition formulas. In his foundational monograph, The Analysis of Linear Partial Differential Operators I, Hörmander defined the principal symbol of a differential operator as a homogeneous function on the cotangent bundle and proved that ellipticity could be characterised in terms of the non-vanishing of this symbol. This framework also allowed for the definition of a full symbol, which includes lower order terms, and enabled the derivation of parametrix constructions for elliptic operators.
Microlocal Analysis
The development of microlocal analysis in the 1970s, spearheaded by mathematicians such as Duistermaat, Hormander, and Guillemin, extended the use of local symbols to study singularities of solutions to partial differential equations. The wave front set of a distribution, for instance, is defined using microlocal notions that rely on the principal symbol of differential and pseudodifferential operators. The local symbol thus plays a crucial role in understanding how singularities propagate along bicharacteristics of the underlying operator.
Definition and Key Concepts
Linear Differential Operators on Manifolds
Let \(M\) be a smooth manifold of dimension \(n\), and let \(P\) be a linear differential operator of order \(m\) acting on smooth functions or sections of a vector bundle over \(M\). In local coordinates \(x = (x^1, \dots, x^n)\), \(P\) can be expressed as \[ P = \sum_{|\alpha|\le m} a_\alpha(x)\,\partial_x^\alpha, \] where \(\alpha = (\alpha_1,\dots,\alpha_n)\) is a multi-index, \(a_\alpha\) are smooth coefficient functions, and \(\partial_x^\alpha = \partial_{x^1}^{\alpha_1}\cdots \partial_{x^n}^{\alpha_n}\).
Principal Symbol
The principal symbol of \(P\) is the homogeneous polynomial of degree \(m\) obtained by retaining only the highest order terms: \[ \sigma_P(x,\xi) = \sum_{|\alpha| = m} a_\alpha(x)\,\xi^\alpha, \] where \(\xi = (\xi_1,\dots,\xi_n)\) is a covector in the cotangent space \(T^*_xM\). The function \(\sigma_P\) is defined on the cotangent bundle \(T^*M\) minus the zero section and is homogeneous of degree \(m\) in \(\xi\). The principal symbol is invariant under change of local coordinates and therefore defines a global smooth function on \(T^*M\setminus \{0\}\).
Local Symbol
While the principal symbol captures the leading order part of a differential operator, the local symbol
Symbol Classes
Hörmander introduced symbol classes \(S^m_{\rho,\delta}\) defined by estimates \[ |\partial_x^\alpha \partial_\xi^\beta \sigma(x,\xi)| \le C_{\alpha,\beta}\,(1+|\xi|)^{m-\rho|\beta| + \delta|\alpha|}, \] with \(0 \le \delta < \rho \le 1\). The most common classes are \(S^m_{1,0}\), where symbols are smooth and satisfy \[ |\partial_x^\alpha \partial_\xi^\beta \sigma(x,\xi)| \le C_{\alpha,\beta}\,(1+|\xi|)^{m-|\beta|}. \] The local symbol of a differential operator of order \(m\) lies in \(S^m_{1,0}\), and the principal symbol lies in \(S^m_{1,0}/S^{m-1}_{1,0}\).
Properties of the Local Symbol
Homogeneity and Scaling
The principal symbol is homogeneous of degree \(m\) in the cotangent variable \(\xi\), meaning that for any \(\lambda > 0\), \[ \sigma_P(x,\lambda\xi) = \lambda^m\,\sigma_P(x,\xi). \] Lower order terms of the local symbol break strict homogeneity, but the scaling behaviour is still controlled by the symbol class estimates.
Behaviour under Composition
For two differential operators \(P\) and \(Q\) of orders \(m\) and \(k\) respectively, the composition \(PQ\) has order \(m+k\). The principal symbol of the composition satisfies \[ \sigma_{PQ}(x,\xi) = \sigma_P(x,\xi)\,\sigma_Q(x,\xi). \] For the full local symbol, the composition formula involves an asymptotic expansion: \[ \tilde\sigma_{PQ}(x,\xi) \sim \sum_{\alpha} \frac{1}{\alpha!}\,\partial_\xi^\alpha \tilde\sigma_P(x,\xi)\,D_x^\alpha \tilde\sigma_Q(x,\xi), \] where \(D_x^\alpha = (-i\partial_x)^\alpha\). This is a special case of the symbolic calculus for pseudodifferential operators.
Adjoint and Complex Conjugate
For a differential operator \(P\) acting on functions on a Riemannian manifold \((M,g)\), the formal adjoint \(P^*\) has principal symbol equal to the complex conjugate of \(\sigma_P\). More precisely, \[ \sigma_{P^*}(x,\xi) = \overline{\sigma_P(x,\xi)}. \] Lower order terms of the local symbol of the adjoint are determined by the metric and the coefficients of \(P\). This property is used in the construction of self‑adjoint elliptic operators.
Examples of Local Symbols
Laplace–Beltrami Operator
The Laplace–Beltrami operator \(\Delta_g\) on a Riemannian manifold has the local symbol \[ \tilde\sigma_{\Delta_g}(x,\xi) = |\xi|_g^2, \] where \(|\xi|_g^2 = g^{ij}(x)\,\xi_i\,\xi_j\). Its principal symbol is the same, reflecting ellipticity: it is non‑vanishing for \(\xi \neq 0\).
Wave Operator
On a Lorentzian manifold \((M,g)\), the d’Alembertian \(\Box_g\) has local symbol \[ \tilde\sigma_{\Box_g}(x,\xi) = -\,g^{ij}(x)\,\xi_i\,\xi_j. \] The principal symbol vanishes on the characteristic set defined by \(g^{ij}\xi_i\xi_j=0\), indicating hyperbolicity.
Schrödinger Operator
For the quantum mechanical Hamiltonian \(H = -\Delta + V(x)\) on \(\mathbb{R}^n\), the local symbol is \[ \tilde\sigma_H(x,\xi) = |\xi|^2 + V(x). \] The principal symbol is \( |\xi|^2 \), which is elliptic, while the potential term represents a lower order perturbation.
General Elliptic Operators
Any linear elliptic operator \(P\) of order \(m\) can be written in local coordinates as \[ \tilde\sigma_P(x,\xi) = \sum_{|\alpha| = m} a_\alpha(x)\,\xi^\alpha + \sum_{|\alpha| < m} a_\alpha(x)\,\xi^\alpha. \] Ellipticity requires that \(\sigma_P(x,\xi) \neq 0\) for all \(x\) and \(\xi \neq 0\).
Applications in Partial Differential Equations
Elliptic Regularity
For an elliptic differential operator \(P\), the invertibility of the principal symbol leads to estimates of the form \[ \|u\|_{H^{s+m}} \le C\bigl(\|Pu\|_{H^s} + \|u\|_{H^{s_0}}\bigr), \] where \(H^s\) denotes the Sobolev space of order \(s\). These estimates imply that solutions to \(Pu=f\) are as smooth as the data \(f\), reflecting the local symbol’s role in establishing regularity.
Propagation of Singularities
In hyperbolic equations, singularities of solutions propagate along bicharacteristic curves defined by the Hamiltonian flow of the principal symbol. The wave front set \(\operatorname{WF}(u)\) of a distribution \(u\) is constrained by the vanishing set of \(\sigma_P\), and the local symbol determines the speed and direction of propagation.
Construction of Parametrices
For elliptic operators, a parametrix \(Q\) is a pseudodifferential operator whose composition with \(P\) differs from the identity by a smoothing operator. The symbol of \(Q\) is constructed as an asymptotic inverse of \(\tilde\sigma_P\), using the local symbol’s leading term as a starting point. This process requires precise knowledge of the local symbol’s expansion.
Spectral Theory
The spectrum of an elliptic operator on a compact manifold can be studied via its symbol. Weyl’s law, for instance, expresses the asymptotic distribution of eigenvalues in terms of integrals involving the principal symbol. The local symbol thus connects differential operator theory with global spectral properties.
Microlocal Analysis and the Local Symbol
Wave Front Set and Symbol Vanishing
The wave front set \(\operatorname{WF}(u)\) is a subset of the cotangent bundle that records singularities of a distribution \(u\) together with their covector directions. A key result is that if \(P\) is a differential operator with symbol \(\sigma_P\) and \(Pu\) is smooth, then \(\operatorname{WF}(u)\) is contained in the characteristic set \(\{\sigma_P=0\}\). The proof relies on the properties of the local symbol under differentiation and Fourier transform.
Fourier Integral Operators
Fourier integral operators (FIOs) generalise pseudodifferential operators by allowing the phase function to vary. The canonical relation associated with an FIO is determined by the Hamiltonian flow of the principal symbol of a related differential operator. The construction and mapping properties of FIOs hinge on the precise behavior of the local symbol near singularities.
Coisotropic Regularity
When studying solutions to \(Pu=0\) near a coisotropic submanifold of \(T^*M\), one examines how the local symbol restricts to that submanifold. Coisotropic regularity results show that solutions exhibit smoothness along certain directions dictated by the local symbol’s invariants.
Numerical Implications of the Local Symbol
Discretisation Error Analysis
Finite difference and finite element discretisations approximate differential operators by discrete analogues. The local symbol of the discretised operator approximates the continuous local symbol up to terms of order \(\Delta x^m\), where \(\Delta x\) is the mesh size. Error estimates often involve the difference between the local symbols of the continuous and discrete operators.
Spectral Methods
In spectral methods, functions are expanded in global basis functions (e.g., Fourier or Chebyshev polynomials). The action of a differential operator on such expansions is mediated by multiplication in the Fourier domain, where the local symbol plays a central role. Accurate computation of the local symbol ensures stability and convergence of spectral schemes.
Preconditioning
In iterative solvers for linear systems derived from discretised PDEs, preconditioners are often designed using approximate inverses of the principal symbol. The local symbol provides the theoretical foundation for selecting effective preconditioners, especially in high‑frequency regimes.
Computational Methods for Local Symbols
Symbolic Computation
Computer algebra systems such as SymPy or MATLAB can symbolically compute local symbols by expanding covector polynomials. For example, in SymPy one can define covector symbols \(\xi_1,\ldots,\xi_n\) and compute \(\tilde\sigma_P\) automatically from the coefficients \(a_\alpha(x)\).
Numerical Approximation of Symbols
In high‑frequency numerical simulations, one may evaluate the local symbol at discrete points in phase space. The estimates from symbol classes guarantee that such evaluations are bounded and smooth, facilitating quadrature methods that approximate integrals over \(T^*M\).
Software Libraries
Numerical libraries such as FEniCS and deal.II provide abstractions for differential operators. While they primarily handle discretised operators, underlying symbolic representations often mimic the structure of local symbols to manage higher‑order terms efficiently.
Future Directions and Open Problems
Non‑Smooth Coefficients
Extending the notion of the local symbol to operators with rough coefficients (e.g., only \(C^{0,\alpha}\)) remains challenging. Current approaches involve microlocal techniques that treat the symbol as a distribution rather than a smooth function, opening avenues for new regularity results.
Quantum Chaos
In quantum chaos, the semiclassical limit of Schrödinger operators with chaotic classical dynamics is investigated via the behavior of the local symbol. Understanding how the lower order terms influence quantum ergodicity and scarring remains an active area of research.
Geometric Inverse Problems
Inverse problems, such as recovering a metric or potential from boundary measurements, rely on uniqueness theorems that involve the principal symbol. Refinements of the local symbol’s role in these problems could lead to improved reconstruction algorithms.
Higher‑Order Pseudodifferential Operators
Recent work explores pseudodifferential operators with exotic symbol classes where the local symbol’s decay in \(\xi\) is slower. These operators appear in the analysis of non‑local PDEs and require new symbolic calculus tools that generalise the classical local symbol concept.
Conclusion
The local symbol of a differential operator encapsulates its complete algebraic structure in the phase space and is fundamental to many areas of analysis, geometry, and physics. Its invariance properties, composition formulas, and scaling behaviour underpin theorems in elliptic regularity, hyperbolic propagation, spectral asymptotics, and microlocal analysis. Ongoing research continues to refine our understanding of the local symbol and to extend its applications to broader classes of operators and manifolds.
External Resources
- Symbol (mathematics) – Wikipedia
- Introduction to Pseudodifferential Operators – Lecture Notes
- Microlocal Analysis – Lecture Notes
No comments yet. Be the first to comment!