03.09.19 · modern-geometry / spin-geometry

Calibrated geometries — Special Lagrangian, associative, coassociative, Cayley

shipped3 tiersLean: partial

Anchor (Master): Harvey-Lawson 1982 *Calibrated geometries* (Acta Mathematica 148); Lawson-Michelsohn §IV.12; Joyce 2007 *Riemannian Holonomy Groups and Calibrated Geometry*; McLean 1998 *Deformations of calibrated submanifolds*

Intuition [Beginner]

A surface in three-space can be wavy or taut. The taut surface — the one a soap film makes when stretched across a wire boundary — has the smallest area among nearby surfaces with the same boundary. This is the prototype of a minimal surface, and finding minimal surfaces in higher dimensions is generally hard: the equations of minimality are second-order, non-linear, and admit many solutions that are not actually area-minimising.

A calibrated geometry is a clever trick that converts an area-minimisation problem into a first-order differential condition, by introducing a special closed differential form on the ambient manifold. The form is chosen so that its value on any unit-volume tangent piece never exceeds one, and the submanifolds whose tangent pieces achieve the value one — the calibrated submanifolds — are automatically area-minimising in their homology class. The proof is one line of integral inequality and one application of Stokes's theorem.

Reginald Harvey and Blaine Lawson introduced this framework in 1982 in their Acta Mathematica paper Calibrated geometries. They showed that the same recipe produces four named geometries on four named ambient manifolds: Special Lagrangian submanifolds inside a Calabi-Yau manifold, associative and coassociative submanifolds inside a manifold, and Cayley submanifolds inside a Spin(7) manifold. All four are calibrated by parallel forms that square out of a parallel spinor on the ambient manifold. This single observation organises four previously disconnected fields of geometry.

Visual [Beginner]

A four-dimensional ambient manifold containing a curved two-dimensional surface, with a small parallelogram of tangent vectors drawn at one point of the surface. A closed differential form is evaluated on this parallelogram; its value is at most the area of the parallelogram. When the value equals the area exactly, the surface is calibrated at that point. Throughout a calibrated submanifold, the equality holds at every tangent point.

A surface inside an ambient manifold, with a tangent parallelogram drawn at one point. A closed differential form lives globally on the ambient manifold; on a unit-area tangent parallelogram it evaluates to at most one. Calibrated submanifolds are those where this value equals one at every point.

The four named calibrations correspond to four named ambient manifolds. Each ambient manifold has its own special holonomy structure that produces the calibrating form as a parallel object.

Worked example [Beginner]

Take complex Euclidean space of complex dimension three, treated as real six-dimensional space with the standard orientation. There is a holomorphic three-form on this space, the determinant of complex coordinates: in the complex setting, which decomposes into a real part and an imaginary part as real differential three-forms.

The real part of is a closed real three-form with the property that on any oriented three-dimensional real subspace of complex three-space, its value (per unit volume of the subspace) lies between minus one and plus one. The subspaces achieving the value plus one are exactly the Special Lagrangian three-planes — the real three-planes that are Lagrangian for the Kähler form and on which the holomorphic volume form takes real positive values.

What this tells us: a three-dimensional real surface in complex three-space whose tangent at every point is a Special Lagrangian plane is automatically a volume-minimiser among nearby three-surfaces with the same boundary. The minimisation reduces to a first-order condition on the tangent plane, no second-order partial differential equations required. This is the simplest of the four Harvey-Lawson examples.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let be a Riemannian manifold of dimension and . The comass of a -form at a point is $$ |\varphi|^\ast(x) := \sup\big{\varphi_x(\xi_1 \wedge \cdots \wedge \xi_p) : \xi_1, \ldots, \xi_p \text{ orthonormal in } T_x M\big}, $$ the supremum of on unit decomposable tangent -vectors. The global comass is .

A calibration of degree is a closed -form on with . The contact set of at is $$ G(\varphi, x) := \big{\xi \in \Lambda^p T_x M : \xi \text{ unit decomposable}, \varphi_x(\xi) = 1\big}, $$ the set of unit decomposable -vectors on which achieves its maximum.

An oriented -dimensional submanifold is calibrated by if at every the oriented unit tangent -vector lies in , i.e., .

The four named calibrations.

(i) Special Lagrangian on a Calabi-Yau -fold : is a closed real -form of comass . The contact set at is the orbit of the real -plane under . Calibrated submanifolds are Special Lagrangian -folds.

(ii) Associative on a manifold : is the parallel three-form, of comass . The contact set is the orbit of the imaginary octonions under . Calibrated submanifolds are associative three-folds.

(iii) Coassociative on a manifold: , the Hodge dual of the three-form, a closed four-form of comass . Calibrated submanifolds are coassociative four-folds.

(iv) Cayley on a Spin(7) manifold : is the parallel Cayley four-form, of comass . The contact set is the orbit under Spin(7) of the standard quaternion . Calibrated submanifolds are Cayley four-folds.

Each calibrating form is parallel — preserved by the Levi-Civita connection — because it arises from a parallel spinor on the ambient manifold via spinor squaring, the construction made explicit through Spin(8) triality and its restrictions to the special-holonomy subgroups of 03.09.13 [Harvey-Lawson 1982]. A calibration is exactly a closed differential form of unit comass; this identifies volume-minimisers with form-tangent submanifolds, the foundational insight of Harvey-Lawson 1982.

Key theorem with proof [Intermediate+]

Theorem (Harvey-Lawson 1982 — fundamental theorem of calibrations). Let be a calibration of degree on a Riemannian manifold , and let be a compact oriented -submanifold calibrated by . Then minimises volume in its relative homology class: for every compact oriented -submanifold with and in , $$ \mathrm{Vol}(N) \leq \mathrm{Vol}(N'). $$

Proof. By the calibration condition and the assumption that is calibrated, $$ \mathrm{Vol}(N) = \int_N \varphi. $$ By Stokes's theorem, since is closed and is the boundary of a -chain in (because in ), $$ \int_N \varphi - \int_{N'} \varphi = \int_W d\varphi = 0. $$ By the comass bound and the elementary inequality that the integral of a -form over an oriented -submanifold is at most the integral of its comass (which is at most the volume), $$ \int_{N'} \varphi \leq \int_{N'} |\varphi|^\ast , d\mathrm{vol}{N'} \leq \int{N'} d\mathrm{vol}{N'} = \mathrm{Vol}(N'). $$ Combining, $$ \mathrm{Vol}(N) = \int_N \varphi = \int{N'} \varphi \leq \mathrm{Vol}(N'). \qquad \square $$

The proof has three moves: (a) the calibration condition turns volume into a -integral on ; (b) the closed-form condition plus Stokes makes ; (c) the comass bound makes . No second-order partial differential equation is needed — the entire proof is one chain of inequalities.

Corollary. Calibrated submanifolds are stable: small variations strictly increase volume unless they remain calibrated. Combined with the parallel-form condition, this means the moduli space of calibrated submanifolds in a fixed cohomology class is a smooth, finite-dimensional manifold whenever it is non-empty (McLean 1998, deformation theory).

Bridge. The construction here builds toward later units of the strand, where the same pattern is taken up at higher structure. The defining pattern appears again in those units in a sharpened form, where the local data is glued or quotiented. Putting these together, the foundational insight is that the data of this unit gives the structural signature that the rest of the strand reads off.

Exercises [Intermediate+]

Lean formalization [Intermediate+]

lean_status: partial — Mathlib lacks the comass functional, the calibration condition, and the special-holonomy structures (Calabi-Yau, , Spin(7)) on which the four named calibrations live. The Lean module declares stub structures for a calibration and a calibrated submanifold; the fundamental theorem is recorded as an axiom pending the upstream pieces.

[object Promise]

The Mathlib gap is the comass functional on differential forms, the calibration class, the four special-holonomy structures, and the fundamental theorem. Once these land, the named calibrations and their calibrated submanifolds become statable as Lean theorems.

Advanced results [Master]

The four named calibrations and their special-holonomy ambients. Each Harvey-Lawson calibration lives on a Riemannian manifold whose holonomy is a special subgroup of stabilising a parallel spinor, and the calibrating form is the spinor-square of that parallel spinor.

Geometry Ambient Holonomy Form Submanifold Dimension of moduli
Special Lagrangian Calabi-Yau -fold (degree ) Lagrangian + phase (McLean)
Associative 7-fold (degree 3) dim of normal twisted Dirac kernel
Coassociative 7-fold (degree 4) (McLean)
Cayley Spin(7) 8-fold Spin(7) (degree 4) dim of normal twisted Dirac kernel

McLean's deformation theorem (1998). For each of the four Harvey-Lawson geometries, the moduli space of calibrated submanifolds in a fixed homology class is a smooth, finite-dimensional manifold near any unobstructed calibrated submanifold, and McLean computed its dimension via index theorems applied to the linearised calibration condition. The Special Lagrangian moduli space has dimension (the first Betti number of the submanifold itself) — McLean's proof reduces the calibration condition to a Hodge-theoretic equation on . The coassociative moduli space has dimension (the dimension of self-dual harmonic 2-forms on ). The associative and Cayley moduli have dimensions computed by twisted-Dirac-operator indices on the normal bundle.

The Strominger-Yau-Zaslow conjecture (1996). Mirror symmetry between Calabi-Yau pairs should manifest as a duality of Special Lagrangian fibrations: and both fibre over a common base by Special Lagrangian tori, and the fibres of one are dual to the fibres of the other (in the sense of -duality). This conjecture, due to Strominger-Yau-Zaslow 1996, places Special Lagrangian geometry at the centre of the mirror-symmetry programme. The conjecture is established for various examples (toric, K3 fibrations) but remains open in general.

Joyce's exotic constructions of and Spin(7) manifolds (1996/2000). Dominic Joyce constructed compact Riemannian 7-manifolds with holonomy and 8-manifolds with Spin(7) holonomy, by resolving orbifolds and with carefully chosen . These were the first compact examples beyond the local Bryant constructions and are the modern testing-ground for associative, coassociative, and Cayley submanifold constructions.

Calibrations beyond the Harvey-Lawson four. Other calibrations exist: the quaternionic Kähler calibrations of degree on quaternionic Kähler manifolds (calibrated by powers of the quaternionic Kähler 4-form); the Cayley-like calibrations on Spin(8) configurations (related to triality acting on octonionic 4-forms). The Harvey-Lawson four remain the most-studied because their ambient holonomy groups admit parallel spinors and so the calibrations are spinor-derived.

Connection to triality. The associative and Cayley calibrations are the spinor-square outputs of triality on Spin(8) restricted to the special-holonomy stabilisers . The associative form on is the orbit of a generic vector under ; the Cayley form on is the orbit of a generic spinor under Spin(7). This is the foundation-of edge from 03.09.13: calibrations built on Spin(7)/ spinor squaring via triality.

Synthesis. This construction generalises the pattern fixed in 03.09.13 (triality on spin(8) and exceptional lie groups via spinors), with the symmetric data replaced by its skew or twisted analogue. Read in the opposite direction, the construction is dual to the metric story: complements and orthogonality are taken with respect to the bilinear datum of this unit, not a metric, and the resulting category of subobjects is the one the rest of the strand classifies. The central insight is that this datum identifies algebra with geometry: functions become vector fields, subspaces become quotients, and invariants become cohomology classes — and that identification is the engine driving every theorem downstream.

Full proof set [Master]

Fundamental theorem of calibrations. Proved in §Key theorem above. The proof is one chain of inequalities: Stokes makes (using closedness and homologousness), and the comass bound makes . The minimisation follows.

Wirtinger inequality. On a Hermitian inner product space with , the form has comass for , with contact set the complex -planes. Proof: at each point , choose a Hermitian orthonormal basis . The form is , and where and is the corresponding -wedge. On any unit decomposable real -vector , the value is bounded by by a Cauchy-Schwarz on each pair, with equality iff is a complex -plane (Federer 1969 §1.8).

Special Lagrangian contact set . Proved in Exercise 6.

Associative contact set . Proved in Exercise 7.

Coassociative contact set. Coassociative 4-planes are exactly the orthogonal complements of associative 3-planes: where is associative. Proof: where the dual takes the orthogonal complement; equality on corresponds to on the orthogonal . The moduli space is then the same homogeneous space as the associative moduli, of real dimension .

Cayley contact set . The Cayley 4-form on has stabiliser inside a 21-dimensional group, and the Cayley contact set is the homogeneous space of real dimension . Cf. Lawson-Michelsohn §IV.12 + Joyce §10.

McLean's deformation theorem (sketch). Linearise the calibration condition at a calibrated submanifold . The linearised calibration constraint is a first-order elliptic equation on the normal bundle , with kernel the infinitesimal calibrated deformations. McLean identifies this kernel with a cohomology group on via Hodge theory: for SL, for coassociative, normal-Dirac-kernel for associative and Cayley. The Sobolev-space implicit-function theorem then promotes infinitesimal deformations to actual moduli. McLean 1998 §3-§5.

Connections [Master]

  • Triality and Spin(8) 03.09.13 — the four calibrating forms arise as spinor squares under the triality-restricted special-holonomy subgroups. The associative form comes from triality on ; the Cayley form from triality on . Foundation-of: calibrations built on Spin(7)/ spinor squaring via triality [conn:415.triality-calibrations, anchor: calibrations built on Spin(7)/G_2 spinor squaring via triality]. This is the consummation of the forward-promise made in 03.09.13's out-edge.

  • Berger holonomy 03.09.18 — each named calibration lives on a manifold with one of Berger's special holonomy groups: for SL, for hyperkähler (the further refinement), for associative/coassociative, for Cayley. The bijection holonomy parallel spinor calibration form is an instance of the Wang-Bryant correspondence. Foundation-of: calibrations require special holonomy structure [conn:427.calibration-special-holonomy, anchor: calibrations require special holonomy structure].

  • Spinor bundle 03.09.05 — the calibrating forms are spinor squares, constructed from a single parallel spinor on the special-holonomy ambient. on Calabi-Yau (with the parallel spinor preserved by ); on ; on Spin(7).

  • Atiyah-Singer / index theorem 03.09.10 — McLean's deformation moduli are dimensions of kernels of normal-bundle elliptic operators (twisted Dirac for associative and Cayley, Hodge-Laplace for SL and coassociative); index-theoretic formulas underwrite the dimension counts.

  • Witten positive-mass theorem 03.09.17 — both calibrated geometry and Witten's positive-mass argument exploit a single closed-form-plus-Stokes manoeuvre. In Witten the form is a divergence built from a harmonic spinor; in Harvey-Lawson the form is a parallel spinor square. The shared technique is Stokes-with-a-pointwise-bound. Recurrence: same Stokes-with-pointwise-bound pattern recurs in spinor positive-mass and calibration arguments.

  • psc obstruction 03.09.16 — the parallel spinor whose square is a calibrating form is a fortiori a nonzero harmonic spinor; on a calibrated Calabi-Yau or or Spin(7) manifold, the spin Dirac kernel is non-empty. The Lichnerowicz argument then forces the scalar curvature to be non-positive somewhere, so calibrated manifolds cannot admit metrics of strictly positive scalar curvature.

We will see in subsequent units the calibrated framework lift to mirror symmetry's SYZ picture and Joyce's compact-holonomy constructions; this builds toward the moduli of calibrated submanifolds and the foreshadowed string-theoretic applications of the next strand. We will later use McLean's deformation-moduli theory to organise the index-theoretic dimensions of calibrated submanifolds, and in the next chapter this pattern recurs in associative G_2 geometry. The calibration philosophy appears again wherever a closed differential form bounds a volume functional. The foundational insight of Harvey-Lawson 1982 is exactly that a closed differential form of unit comass identifies its tangent submanifolds with volume-minimisers — this is precisely the same statement as the Stokes-with-comass argument. Putting these together shows that each calibrated geometry is an instance of one universal definition. The bridge between calibration data and special holonomy is the parallel-spinor square, and each named geometry is precisely the corresponding spinor structure on a Wang-special-holonomy manifold.

Historical & philosophical context [Master]

Reginald Harvey and Blaine Lawson's 1982 Calibrated geometries (Acta Mathematica 148, 47–157) is the founding paper of the field. Lawson is co-originator. The paper opens with a clean abstract structural definition of calibration — a closed differential form of unit comass — and treats Special Lagrangian, associative/coassociative, and Cayley calibrations in parallel as four named instances of one structure. The fundamental theorem (calibrated submanifolds are volume-minimisers) is proved in §I via the one-line Stokes-with-comass argument reproduced above. Parts II, III, IV then develop the four named geometries, each with a self-contained treatment of the contact set, the calibrated submanifolds, and the local examples.

Harvey and Lawson's framing in the introduction is structural: they note that the Wirtinger inequality (1936) and the Federer-Fleming theorem on complex submanifolds had been known for decades, but had never been recognised as instances of a general phenomenon. The conceptual contribution of Calibrated geometries is the unifying definition itself — the recognition that one functional analytic condition (comass plus closed) produces, in different ambient settings, all the volume-minimising geometries of interest. As Harvey-Lawson put it in the preface: the four named geometries "exhibit a remarkable parallelism that could not be coincidence." The paper articulates that parallelism by reducing it to a single structural definition.

The historical moment matters. By 1982, Calabi-Yau manifolds were well-developed in algebraic geometry (Yau's 1977 proof of the Calabi conjecture had landed five years prior); holonomy was newly proved to exist on local examples (Bryant's 1987 result was still in progress); Spin(7) holonomy was even more nascent. Harvey-Lawson's paper preceded the major compact and Spin(7) constructions (Joyce 1996/2000) by over a decade. They were defining the geometries on the belief that compact examples would exist — a belief vindicated when Joyce's constructions delivered them. The paper is therefore the foundation of an entire research programme that took shape only after its publication, with the calibrated-submanifold fundamental theorem providing the central technique throughout.

Marcel Berger's 1955 holonomy classification (see 03.09.18) had identified as the special holonomy groups stabilising a parallel spinor; Wang's 1989 Parallel spinors later made this precise as a bijection between holonomy groups and parallel-spinor counts. Harvey-Lawson's paper sits between these two: Berger gave the list of holonomies; Harvey-Lawson explained what each of those holonomies does, geometrically — it produces a calibration. The downstream development (Joyce, McLean, Strominger-Yau-Zaslow, Donaldson-Thomas) all takes the Harvey-Lawson calibrations as primary objects.

The paper is also unusual in its narrative completeness. Acta Mathematica 148 is 110 pages; the four named geometries each receive a chapter of treatment; the moduli theory of calibrated submanifolds is developed as far as it could be without the deformation analysis that McLean would supply sixteen years later. The paper reads as a complete theory rather than a research announcement — Lawson's mature mathematical-textbook voice is on full display, the same voice that animates Lawson-Michelsohn 1989. A reader of Lawson-Michelsohn §IV.12 hears that voice condensed; a reader of Harvey-Lawson 1982 hears it at full development.

Bibliography [Master]

  • Harvey, R. & Lawson, H. B., "Calibrated geometries", Acta Mathematica 148 (1982), 47–157. (Originating paper; Lawson is co-originator.)
  • Lawson, H. B. & Michelsohn, M.-L., Spin Geometry, Princeton University Press, 1989. §IV.12.
  • McLean, R. C., "Deformations of calibrated submanifolds", Communications in Analysis and Geometry 6 (1998), 705–747.
  • Joyce, D. D., Compact Manifolds with Special Holonomy, Oxford University Press, 2000.
  • Joyce, D. D., Riemannian Holonomy Groups and Calibrated Geometry, Oxford University Press, 2007.
  • Strominger, A., Yau, S.-T. & Zaslow, E., "Mirror symmetry is T-duality", Nuclear Physics B 479 (1996), 243–259.
  • Wirtinger, W., "Eine Determinantenidentität und ihre Anwendung auf analytische Gebilde", Monatshefte für Mathematik 44 (1936), 343–365. (The prototype Wirtinger inequality, predating Harvey-Lawson by 46 years.)
  • Federer, H., Geometric Measure Theory, Springer, 1969. §1.8.
  • Bryant, R. L., "Metrics with exceptional holonomy", Annals of Mathematics 126 (1987), 525–576.