lecture01
lecture01
MATH 136
Introduction
1.1. Differential geometry deals with geometric objects called manifolds. Mani-
folds can be described intrinsically. It is convenient however to look first at manifolds
embedded in a Euclidean space Rn like our space R3 . Examples are one dimensional
manifolds called curves or two dimensional manifolds called surfaces or hypersurfaces
given as a root of a function f . The 3-sphere x21 + x21 + x23 + x24 = 1 for example
is a 3-dimensional manifold embedded in 4-dimensional space. Since we live in a 3-
dimensional space, we are at first mostly interested in one or two dimensional manifolds
in R3 . These objects are called curves or surfaces.
1.2. A curve is given as a parametrization r(t) = [x(t), y(t), z(t)] where t ∈ [a, b]
is the parameter domain and x(t), y(t), z(t) are functions of one variable. In two di-
mensions, some curves can also be written implicitly as a level curve like for ex-
ample the circle x2 + y 2 = 1. A surface is given as a parametrization r(u, v) =
[x(u, v), y(u, v), z(u, v)] where (u, v) is in some domain R ⊂ R2 in the uv-plane. Some
surfaces can be given as level surfaces f (x, y, z) = 0. Some curves can be given as
intersection of level surfaces f (x, y, z) = 0, g(x, y, z) = 0.
Rb ′
1.3. We RR are interested in global quantities like arc length a |r (t)| dt or surface
area R |ru ×rv | dudv, where × is the cross product as well as local quantities like
curvature κ(t) = |r′ (t) × r′′ (t)|/|r′ (t)|3 or torsion τ (t) = det[r′ (t), r′′ (t), r′′′ (t)]/|r′ ×
r′′ |2 for curves. For a surface, the curvature of a point can be defined as K(p) =
limr→0 3 2πr−|S r (p)|)
πr3
, where |Sr (p)| is the length of the wave front Sr (p) of points on the
surface in distance r from p which is a circle for small r. This is an intrinsic definition
not making use of the embedding of the surface in an ambient space. It even makes
sense on non-smooth surfaces, like polyhedra.
1.4. Curvature plays an important role in differential geometry. We will define it dif-
ferently later in the course and verify that it is independent of the embedding in space.
This is the Theorema egregium, the ”great theorem” of Gauss from 1827. Riemann-
ian geometry, the idea of doing geometry on a manifold without having to embed it into
an ambient emerged in an inaugural lecture of Riemann in 1854. The theory is used
heavily in Einstein’s 1915 theory of general relativity to which Schwarzschild found a
black hole solution in 1916. A 100 years later, gravitational waves produced by black
hole mergers have been observed.
Differential Geometry
Figure 1. To the left we see a curve called Euler curve, to the right, we
see a visualization of the 3 dimensional sphere x2 + y 2 + z 2 + w2 = 1 in
R4 . It is foliated by 2 dimensional flat tori. We can not embed flat tori
in R3 but we can in R4 .
1.5. The focus in differential geometry is to investigate relations between local quanti-
ties and global properties like for example to integrate up curvature. We are especially
interested in quantities that do not depend on the metric, like the Euler character-
istic. Here are examples which will appear early in this course: for a planar closed
curve in R2 one can define the signedR curvature as κ(t) = (r′ (t) × r′′ (t))/|r′ (t)|3 . We
will then see the Hopf Umlaufsatz C κ(t) dr(t) = 2π. For a two-dimensional surface
with gRRholes of Euler characteristic χ(G) = 2 − 2g, one has the Gauss-Bonnet theo-
rem R K(x) dV (x) = 2πχ(G). We also want to understand geodesics, curves that
locally minimize length. One can start geodesics into any direction v/|v| and let it run
for a distance |v|. This produces the exponential map expp , a map from the tangent
space Tp M of a point p to the manifold. If Sr is the sphere of radius r in R2 , then
the image Wr (t) = expp (Sr ) is called the wave front at p. These waves can become
complicated for large r. This can also be studied on polyhedra. We expect wave fronts
to become dense in the manifold, except for very special cases like the round sphere.
1.6. Differential geometry then extends curve and surface theory to arbitrary dimen-
sions. One study then so called Riemannian manifolds or pseudo-Riemannian
manifolds which appear in physics. There is an intrinsic geometry but also interest
when manifolds M are embedded in larger manifolds M ′ . In general relativity for ex-
ample, space is a 3-dimensional manifold embedded in a four dimensional space-time
manifold M ′ . The above formulation of curves or surfaces dealt with embeddings
of one or two dimensional manifolds in Euclidean 3-manifold M ′ . One can use the
exponential map to define sectional curvature and to use it to define a curvature
tensor or scalar curvature. The extrema of the functional that gives the total scalar
curvature are the Einstein equations. General relativity studies solutions of these
equations as they tell how matter bends space. The geodesic equations then tell, how
matter moves in this space.
Oliver Knill, knill@math.harvard.edu, Math 136, Fall, 2024