SYSTEMATIC MATHEMATICS

Geometric Measure Theory & Minimal Surfaces

How can a curve have zero planar area but positive length, how can a derivative be a measure concentrated on a jump, how can an oriented surface converge after its parametrization breaks, why can opposite current sheets cancel while varifold sheets add, and when does a weak stationary object become a smooth minimal hypersurface? This Stage 5 doctoral-specialization core answers those questions in eighty-four visible bilingual chapters across twelve content-sized units of lengths 7, 7, 7, 7, 8, 7, 7, 8, 7, 6, 7, and 6. It constructs Radon and Hausdorff measures, densities, Frostman bounds, Lipschitz differentiation, Jacobians, area and coarea, rectifiability, approximate tangents, projections, tangent measures, beta numbers, BV derivatives, finite-perimeter sets, reduced boundaries, and Gauss–Green. It then builds differential forms, currents, mass, boundary, pushforward, normal and integral compactness, slicing, isoperimetric fillings, the weak Plateau method, flat distance, varifolds, first variation, generalized mean curvature, monotonicity, Allard compactness and regularity, minimal-graph equations, stability, Jacobi fields, excess decay, curvature estimates, tangent cones, singular strata, and dimension reduction. Final chapters treat calibrations, competing Plateau formulations, prescribed mean curvature, capillarity, anisotropic Wulff energies, Brakke and level-set flows, and six independently reproducible original research dossiers. Every chapter defines its symbols in words, derives one exact mechanism, solves a concrete example, identifies the nearest failure, and separates finite-resolution evidence from a continuum theorem. This course provides doctoral research preparation in Euclidean geometric measure theory, codimension-one minimal hypersurfaces, selected higher-codimension current phenomena, and weak mean-curvature motion; it does not claim complete coverage of metric-space GMT, Almgren multiple-valued theory, every coefficient group, every boundary regularity theorem, arbitrary geometric flows, or a universal postdoctoral syllabus.

Before this course: Completed Measure & Lebesgue Integration; Differential Geometry & Manifolds; Differential Forms, de Rham Cohomology & Hodge Theory; Sobolev Spaces, Distributions & Weak PDE; Calculus of Variations; and Nonlinear Elliptic & Parabolic Regularity, together with their prerequisite routes in real analysis, topology, linear algebra, vector calculus, and functional analysis. Learners must already be able to use Radon measures, weak-star compactness, distributional integration by parts, differential forms and Stokes, tangent spaces and curvature, Sobolev compactness, direct minimization, elliptic regularity, and weak geometric-flow language. No prior unified course in Hausdorff measure, rectifiability, BV perimeter, currents, varifolds, Allard theory, Plateau currents, singular-cone stratification, or Brakke flow is assumed.

COURSE FACTSLevel, chapters, units, prerequisite, and outcome
Chapter 1

Irregular sets require dimension-aware size rather than one universal notion of volume

Objective: Why is “the segment has measure zero” incomplete?

Lebesgue volume correctly measures full-dimensional regions but assigns zero to smooth curves in the plane and smooth surfaces in space. Geometric measure theory therefore asks both how large a set is and in which dimension it carries mass. Begin by declaring the ambient metric space, dimension being measured, measure normalization, orientation if any, and whether the conclusion concerns every point, almost every point, a regular set, or a singular set.

Read every geometric symbol through its definition: For E⊂R^n, L^n(E) denotes n-dimensional Lebesgue measure. A dimension-indexed family H^s will later measure s-dimensional size. The ambient dimension n, measured dimension s, metric, and normalization are separate inputs. The exact conclusion established here is: A line segment of length ℓ in R² has L² measure zero but should receive one-dimensional size ℓ under the standard H¹ normalization; no contradiction occurs because the two measures answer different dimensional questions.

The proof route is auditable: 1. Cover the segment by a rectangle of length ℓ and width ε; its planar area is ℓε. 2. Let ε↓0 to make the area of an open cover arbitrarily small, proving L² of the segment is zero. 3. Measure along the segment parameter t↦a+tv for 0≤t≤ℓ; unit speed accumulates one-dimensional length ∫_0^ℓ1dt=ℓ. At each limit, retain the uniform mass or variation bound, compactness topology, lower-semicontinuity step, multiplicity, boundary, and hypothesis preventing cancellation or diffuse loss.

Reconstruct “A segment in R² has length three and is covered by a rectangle of width 1/100. Compute the rectangle area and state the segment’s planar measure and standard one-dimensional measure.” before invoking a general theorem. Then test the nearest failure: A set can have zero s-dimensional measure while having dimension s, or infinite H^s while carrying finite H^t for another t. Dimension and measure value must never be collapsed into one number. This separates a picture from a measure statement, a stationary object from a minimizer, and compactness from regularity.

A line segment of length ℓ in R² has L² measure zero but should receive one-dimensional size ℓ under the standard H¹ normalization; no contradiction occurs because the two measures answer different dimensional questions.

Cover the segment by a rectangle of length ℓ and width ε; its planar area is ℓε.

Let ε↓0 to make the area of an open cover arbitrarily small, proving L² of the segment is zero.

Measure along the segment parameter t↦a+tv for 0≤t≤ℓ; unit speed accumulates one-dimensional length ∫_0^ℓ1dt=ℓ.

A segment in R² has length three and is covered by a rectangle of width 1/100. Compute the rectangle area and state the segment’s planar measure and standard one-dimensional measure.

  1. The covering rectangle area is 3·(1/100)=3/100.
  2. Allowing the width to tend to zero shows the segment has planar Lebesgue measure zero.
  3. Under standard length normalization, its H¹ measure equals its length, namely three.

Result: The finite cover has area 3/100; the segment itself has L² measure zero and H¹ measure three.