A family of techniques that isolate and study singularities of distributions simultaneously in space (position) and frequency (Fourier variables), typically by localizing in phase space with pseudodifferential operators, wave packets, or FBI transforms to reveal the wavefront set and fine regularity/propagation properties.

Principle

Singularities are best understood not only by their location but by the directions (cotangent/frequency) in which they fail to be smooth; microlocalization creates localized probes in phase space that respect the uncertainty principle and track how PDEs transport singularities along characteristic directions.

Demonstration

For the wave equation on a manifold, microlocalization using pseudodifferential cutoffs shows that the wavefront set of a solution propagates along null bicharacteristics; applying a microlocal parametrix yields precise regularity improvement away from characteristic directions.

Misapplication

Attempting arbitrarily sharp simultaneous localization in position and frequency in violation of the uncertainty principle, or using naive truncation in the Fourier variable without regard to smooth cutoffs and microlocal calculus, leading to spurious conclusions about singular support.

Consequence

Microlocalization produces exact statements about propagation of singularities, enables construction of parametrices and elliptic regularity results in restricted directions, and decomposes problems into phase-space pieces amenable to symbolic calculus and semiclassical analysis.

Reversal

Classical localization in position alone (local cutoff in x) or global spectral methods that ignore directional frequency information; these recover only coarse singular support information and miss directional propagation phenomena.

Boundary

Operates for distributions, hyperfunctions, and tempered distributions on manifolds or R^n with a cotangent phase space; it does not apply in meaningful form to purely finite-dimensional algebraic objects with no notion of frequency, nor to problems where no Fourier or cotangent description exists.

Semantic Tension

Tension arises between microlocal (phase-space directional) descriptions and purely spatial regularity: a function may be smooth in x but have directional frequency singularities that microlocal analysis detects; conversely coarse spectral methods may conflate distinct microlocal behaviors.

Synthesis

Microlocalization refines localization by adding directional/frequency information to spatial position; by decomposing distributions in phase space it turns global PDE and regularity questions into local analyses along characteristic directions, reconciling local smoothness with directional singularities.