Étale fundamental group explained
The étale or algebraic fundamental group is an analogue in algebraic geometry, for schemes, of the usual fundamental group of topological spaces.
Topological analogue/informal discussion
In algebraic topology, the fundamental group
of a pointed topological space
is defined as the
group of homotopy classes of loops based at
. This definition works well for spaces such as real and complex
manifolds, but gives undesirable results for an
algebraic variety with the
Zariski topology.
In the classification of covering spaces, it is shown that the fundamental group is exactly the group of deck transformations of the universal covering space. This is more promising: finite étale morphisms of algebraic varieties are the appropriate analogue of covering spaces of topological spaces. Unfortunately, an algebraic variety
often fails to have a "universal cover" that is finite over
, so one must consider the entire category of finite étale coverings of
. One can then define the étale fundamental group as an
inverse limit of finite
automorphism groups.
Formal definition
Let
be a connected and locally
noetherian scheme, let
be a geometric point of
and let
be the category of pairs
such that
is a
finite étale morphism from a scheme
Morphisms
in this category are morphisms
as
schemes over
This category has a
natural functor to the category of sets, namely the functor:
F(Y)=\operatorname{Hom}X(x,Y);
geometrically this is the fiber of
over
and abstractly it is the
Yoneda functor represented by
in the category of schemes over
. The functor
is typically not representable in
; however, it is pro-representable in
, in fact by Galois covers of
. This means that we have a projective system
in
, indexed by a
directed set
where the
are
Galois covers of
, i.e., finite étale schemes over
such that
\#\operatorname{Aut}X(Xi)=\operatorname{deg}(Xi/X)
.
[1] It also means that we have given an isomorphism of functors:
F(Y)=\varinjlimi\operatorname{Hom}C(Xi,Y)
.In particular, we have a marked point
of the projective system.
For two such
the map
induces a group homomorphism
\operatorname{Aut}X(Xj)\to\operatorname{Aut}X(Xi)
which produces a projective system of automorphism groups from the projective system
. We then make the following definition: the
étale fundamental group
of
at
is the inverse limit:
\pi1(X,x)=\varprojlimi{\operatorname{Aut}}X(Xi),
with the inverse limit topology.
The functor
is now a functor from
to the category of finite and continuous
-sets and establishes an
equivalence of categories between
and the category of finite and continuous
-sets.
Examples and theorems
The most basic example of is
\pi1(\operatorname{Spec}k)
, the étale fundamental group of a
field
. Essentially by definition, the fundamental group of
can be shown to be isomorphic to the
absolute Galois group \operatorname{Gal}(ksep/k)
. More precisely, the choice of a geometric point of
is equivalent to giving a separably closed extension field
, and the étale fundamental group with respect to that base point identifies with the Galois group
. This interpretation of the Galois group is known as
Grothendieck's Galois theory.
More generally, for any geometrically connected variety
over a field
(i.e.,
is such that
is connected) there is an
exact sequence of
profinite groups:
1\to
,\overline{x})\to\pi1(X,\overline{x})\to\operatorname{Gal}(ksep/k)\to1.
Schemes over a field of characteristic zero
For a scheme
that is of finite type over
, the complex numbers, there is a close relation between the étale fundamental group of
and the usual, topological, fundamental group of
, the complex analytic space attached to
. The algebraic fundamental group, as it is typically called in this case, is the profinite completion of
. This is a consequence of the Riemann existence theorem, which says that all finite étale coverings of
stem from ones of
. In particular, as the fundamental group of smooth curves over
(i.e., open Riemann surfaces) is well understood; this determines the algebraic fundamental group. More generally, the fundamental group of a proper scheme over any algebraically closed field of characteristic zero is known, because an extension of algebraically closed fields induces isomorphic fundamental groups.Schemes over a field of positive characteristic and the tame fundamental group
For an algebraically closed field
of positive characteristic, the results are different, since Artin–Schreier coverings exist in this situation. For example, the fundamental group of the affine line
is not topologically finitely generated. The tame fundamental group
of some scheme U
is a quotient of the usual fundamental group of
which takes into account only covers that are tamely ramified along
, where
is some compactification and
is the complement of
in
. For example, the tame fundamental group of the affine line is zero.Affine schemes over a field of characteristic p
It turns out that every affine scheme
is a
-space, in the sense that the etale homotopy type of
is entirely determined by its etale homotopy group.
[2] Note
where
is a geometric point.
Further topics
From a category-theoretic point of view, the fundamental group is a functor:
→ .The inverse Galois problem asks what groups can arise as fundamental groups (or Galois groups of field extensions). Anabelian geometry, for example Grothendieck's section conjecture, seeks to identify classes of varieties which are determined by their fundamental groups.
studies higher étale homotopy groups by means of the étale homotopy type of a scheme.
The pro-étale fundamental group
have introduced a variant of the étale fundamental group called the pro-étale fundamental group. It is constructed by considering, instead of finite étale covers, maps which are both étale and satisfy the valuative criterion of properness. For geometrically unibranch schemes (e.g., normal schemes), the two approaches agree, but in general the pro-étale fundamental group is a finer invariant: its profinite completion is the étale fundamental group.
See also
Notes
- J. S. Milne, Lectures on Étale Cohomology, version 2.21: 26-27
- Achinger. Piotr. November 2017. Wild ramification and K(pi, 1) spaces. Inventiones Mathematicae. 210. 2. 453–499. 10.1007/s00222-017-0733-5. 1701.03197. 119146164. 0020-9910.