1 of 51

differential geometers suck at naming things

alex pawełko

2 of 51

“Differential geometry is the subject of things invariant under change-of-notation”

-Apocryphal, popularized by John M. Lee

3 of 51

4 of 51

5 of 51

6 of 51

7 of 51

8 of 51

9 of 51

10 of 51

11 of 51

12 of 51

13 of 51

14 of 51

15 of 51

16 of 51

we’re really bad at picking names

17 of 51

we’re really bad at picking names

18 of 51

we’re really bad at picking names

19 of 51

we’re really bad at picking names

20 of 51

we’re really bad at picking names

21 of 51

we’re really bad at picking names

22 of 51

we’re really bad at picking names

23 of 51

we’re really bad at picking names

24 of 51

we’re really bad at picking names

25 of 51

we’re really bad at picking names

26 of 51

we’re really bad at picking names

27 of 51

28 of 51

Theorem (Berger’s Holonomy Theorem, 1955)

Let (M, g) be an irreducible nonsymmetric Riemannian manifold. Then Hol(g) is one of the following groups:

29 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

30 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

31 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

32 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

33 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

34 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

35 of 51

group

name

dimension

SO(n)

generic case (oriented Riemannian)

any

U(m)

Kaehler (torsion-free Hermitian)

n = 2m is even

SU(m)

Calabi-Yau (torsion-free complex-oriented Hermitian)

n = 2m is even

Sp(m)

hyper-Kaehler (torsion-free quaternionic Hermitian)

n = 4m

Sp(m)Sp(1)

Quaternionic Kaehler

n = 4m

G2

(torsion-free) G2

n = 7

Spin(7)

(torsion-free) Spin(7)

n = 8

36 of 51

these pde are really totally ass

37 of 51

Let’s look at U(m)-structures:

By the holonomy principle, get linear algebraic objects:�

g, J, ω, NJ, θ

38 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

39 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

40 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

41 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

42 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

43 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

44 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

45 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

46 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

47 of 51

Name

PDE

Kaehler

Nearly Kaehler

Almost Kaehler

Hermitian semi-Kaehler

Locally conformally Kaehler

quasi-Kaehler

Hermitian

coalmost Kaehler

semi-Kaehler

Locally conformally almost Kaehler

48 of 51

There are more… (whose PDE are really annoying to write down)

pseudo-Kaehler

paraKaehler

nearly pseudo-Kaehler

almost pseudo-Kaehler (this is the same PDE as almost Kaehler with an extra linear algebraic condition)

nearly paraKaehler

49 of 51

And the ultimate paper

(which i actually read)

50 of 51

51 of 51

You understand everything in this equation�(there are 3 implicit sums, 6 implicit derivatives, and many many omitted isomorphisms we are assuming are =)

Gamma^k_{ij}d/dx^k = \nab_i d/dx^j