Вы находитесь на странице: 1из 50

ECohomology

From Wikipedia, the free encyclopedia

Jump to navigationJump to search

In mathematics, specifically in homology theory and algebraic


topology, cohomology is a general term for a sequence of abelian
groups associated to a topological space, often defined from a cochain
complex. Cohomology can be viewed as a method of assigning richer algebraic
invariants to a space than homology. Some versions of cohomology arise by
dualizing the construction of homology. In other words, cochains
are functions on the group of chains in homology theory.
From its beginning in topology, this idea became a dominant method in the
mathematics of the second half of the twentieth century. From the initial idea of
homology as a method of constructing algebraic invariants of topological
spaces, the range of applications of homology and cohomology theories has
spread throughout geometry and algebra. The terminology tends to hide the fact
that cohomology, a contravariant theory, is more natural than homology in many
applications. At a basic level, this has to do with functions and pullbacks in
geometric situations: given spaces X and Y, and some kind of function F on Y,
for any mapping f : X → Y, composition with f gives rise to a function F ∘ f on X.
The most important cohomology theories have a product, the cup product,
which gives them a ring structure. Because of this feature, cohomology is
usually a stronger invariant than homology.

Contents

 1Singular cohomology

 2Examples

 3The diagonal

 4Poincaré duality

 5Characteristic classes

 6Eilenberg–MacLane spaces

 7Cap product

 8History, to the birth of singular cohomology

 9Sheaf cohomology

 10Cohomology of varieties

 11Axioms and generalized cohomology theories


 12Other cohomology theories

 13See also

 14Notes

 15References

Singular cohomology[edit]
Singular cohomology is a powerful invariant in topology, associating a graded-
commutative ring to any topological space. Every continuous
map f: X → Y determines a homomorphism from the cohomology ring of Y to
that of X; this puts strong restrictions on the possible maps from X to Y. Unlike
more subtle invariants such as homotopy groups, the cohomology ring tends to
be computable in practice for spaces of interest.
For a topological space X, the definition of singular cohomology starts with
the singular chain complex:[1]
By definition, the singular homology of X is the homology of this chain
complex (the kernel of one homomorphism modulo the image of the
previous one). In more detail, Ci is the free abelian group on the set of
continuous maps from the standard i-simplex to X (called "singular i-
simplices in X"), and ∂i is the ith boundary homomorphism. The groups Ci are
zero for i negative.
Now fix an abelian group A, and replace each group Ci by its dual
group  and  by its dual homomorphism
This has the effect of "reversing all the arrows" of the original complex,
leaving a cochain complex
For an integer i, the ith cohomology group of X with coefficients
in A is defined to be ker(di)/im(di−1) and denoted by Hi(X, A). The
group Hi(X, A) is zero for i negative. The elements of  are
called singular i-cochains with coefficients in A. (Equivalently, an i-
cochain on X can be identified with a function from the set of
singular i-simplices in X to A.) Elements of ker(d) and im(d) are
called cocycles and coboundaries, respectively, while elements of
ker(d)/im(d) = Hi(X, A) are called cohomology classes (because
they are equivalence classes of cocycles).
In what follows, the coefficient group A is sometimes not written. It is
common to take A to be a commutative ring R; then the cohomology
groups are R-modules. A standard choice is the ring Z of integers.
Some of the formal properties of cohomology are only minor variants
of the properties of homology:
 A continuous map  determines a pushforward homomorphism  on
homology and a pullback homomorphism  on cohomology. This makes
cohomology into a contravariant functor from topological spaces to
abelian groups (or R-modules).
 Two homotopic maps from X to Y induce the same homomorphism on
cohomology (just as on homology).

 The Mayer–Vietoris sequence is an important computational tool in


cohomology, as in homology. Note that the boundary homomorphism
increases (rather than decreases) degree in cohomology. That is, if a
space X is the union of open subsets U and V, then there is a long
exact sequence:

 There are relative cohomology groups Hi(X,Y;A) for


any subspace Y of a space X. They are related to the usual
cohomology groups by a long exact sequence:

 The universal coefficient theorem describes cohomology in


terms of homology, using Ext groups. Namely, there is a short
exact sequence
A related statement is that for a field F, Hi(X,F) is precisely the dual space of
the vector space Hi(X,F).

 If X is a topological manifold or a CW complex, then the


cohomology groups Hi(X,A) are zero for i greater than
the dimension of X.[2] If X is a compact manifold
(possibly with boundary), or a CW complex with finitely
many cells in each dimension, and R is a
commutative Noetherian ring, then the R-
module Hi(X,R) is finitely generated for each i.[3]
On the other hand, cohomology has a crucial structure
that homology does not: for any topological
space X and commutative ring R, there is a bilinear
map, called the cup product:
defined by an explicit formula on singular cochains.
The product of cohomology classes u and v is
written as u ∪ v or simply as uv. This product
makes the direct sum
into a graded ring, called the cohomology
ring of X. It is graded-commutative in the
sense that:[4]
For any continuous map  the pullback  is a
homomorphism of graded R-algebras. It
follows that if two spaces are homotopy
equivalent, then their cohomology rings are
isomorphic.
Here are some of the geometric
interpretations of the cup product. In what
follows, manifolds are understood to be
without boundary, unless stated otherwise.
A closed manifold means a compact
manifold (without boundary), whereas
a closed submanifold N of a
manifold M means a submanifold that is
a closed subset of M, not necessarily
compact (although N is automatically
compact if M is).

 Let X be a closed oriented manifold of
dimension n. Then Poincaré duality gives
an isomorphism HiX ≅ Hn−iX. As a result, a
closed oriented
submanifold S of codimension i in X determi
nes a cohomology class in HiX, called [S].
In these terms, the cup product describes
the intersection of submanifolds. Namely,
if S and T are submanifolds of
codimension i and j that
intersect transversely, then
where the intersection S ∩ T is a submanifold of codimension i + j, with an
orientation determined by the orientations of S, T, and X. In the case of smooth
manifolds, if S and T do not intersect transversely, this formula can still be used
to compute the cup product [S][T], by perturbing S or T to make the intersection
transverse.
More generally, without assuming that X has an orientation, a closed
submanifold of X with an orientation on its normal bundle determines a
cohomology class on X. If X is a noncompact manifold, then a closed
submanifold (not necessarily compact) determines a cohomology class on X. In
both cases, the cup product can again be described in terms of intersections of
submanifolds.
Note that Thom constructed an integral cohomology class of degree 7 on a
smooth 14-manifold that is not the class of any smooth submanifold.[5] On the
other hand, he showed that every integral cohomology class of positive degree
on a smooth manifold has a positive multiple that is the class of a smooth
submanifold.[6] Also, every integral cohomology class on a manifold can be
represented by a "pseudomanifold", that is, a simplicial complex that is a
manifold outside a closed subset of codimension at least 2.

 For a smooth
manifold X, de Rham's
theorem says that the
singular cohomology
of X with real coefficients is
isomorphic to the de Rham
cohomology of X, defined
using differential forms. The
cup product corresponds to
the product of differential
forms. This interpretation
has the advantage that the
product on differential forms
is graded-commutative,
whereas the product on
singular cochains is only
graded-commutative up
to chain homotopy. In fact,
it is impossible to modify
the definition of singular
cochains with coefficients in
the integers Z or in Z/p for a
prime number p to make
the product graded-
commutative on the nose.
The failure of graded-
commutativity at the
cochain level leads to
the Steenrod operations on
mod p cohomology.
Very informally, for any
topological space X,
elements of HiX can be
thought of as represented by
codimension-i subspaces
of X that can move freely
on X. For example, one way
to define an element of HiX is
to give a continuous
map f from X to a
manifold M and a closed
codimension-i submanifold N 
of M with an orientation on
the normal bundle.
Informally, one thinks of the
resulting class  as lying on
the subspace  of X; this is
justified in that the
class  restricts to zero in the
cohomology of the open
subset  The cohomology
class  can move freely
on X in the sense
that N could be replaced by
any continuous deformation
of N inside M.

Examples[edit]
In what follows, cohomology
is taken with coefficients in
the integers Z, unless stated
otherwise.

 The cohomology ring of a


point is the ring Z in degree
0. By homotopy invariance,
this is also the cohomology
ring of
any contractible space,
such as Euclidean
space Rn.

The first cohomology group of


the 2-dimensional torus has a
basis given by the classes of the
two circles shown.

For a positive integer n, the


cohomology ring of
the sphere Sn is Z[x]/(x2)
(the quotient ring of
a polynomial ring by the
given ideal), with x in
degree n. In terms of
Poincaré duality as
above, x is the class of a
point on the sphere.

 The cohomology ring of


the torus (S1)n is the exterior
algebra over Z on n generat
ors in degree 1.[7] For
example, let P denote a
point in the circle S1,
and Q the point (P,P) in the
2-dimensional torus (S1)2.
Then the cohomology of
(S1)2 has a basis as
a free Z-module of the form:
the element 1 in degree
0, x := [P × S1] and y :=
[S1 × P] in degree 1,
and xy = [Q] in degree 2.
(Implicitly, orientations of
the torus and of the two
circles have been fixed
here.) Note that yx = −xy =
−[Q], by graded-
commutativity.
 More generally, let R be a
commutative ring, and
let X and Y be any
topological spaces such
that H*(X,R) is a finitely
generated free R-module in
each degree. (No
assumption is needed
on Y.) Then the Künneth
formula gives that the
cohomology ring of
the product space X × Y is
a tensor product of R-
algebras:[8]

 The cohomology ring


of real projective
space RPn with Z/2
coefficients is Z/2[x]/
(xn+1), with x in degree 1.
[9]
 Here x is the class of
a hyperplane RPn−1 in R
Pn; this makes sense
even though RPj is not
orientable for j even
and positive, because
Poincaré duality
with Z/2 coefficients
works for arbitrary
manifolds.
With integer coefficients, the answer is a bit more complicated. The Z-
cohomology of RP2a has an element y of degree 2 such that the whole
cohomology is the direct sum of a copy of Z spanned by the element 1 in
degree 0 together with copies of Z/2 spanned by the elements yi for i=1,...,a.
The Z-cohomology of RP2a+1 is the same together with an extra copy of Z in
degree 2a+1.[10]

 The cohomology
ring of complex
projective
space CPn is Z[x]/(xn
+1
), with x in degree
2.[9] Here x is the
class of a
hyperplane CPn−1 in 
CPn. More
generally, xj is the
class of a linear
subspace CPn−j in C
Pn.
 The cohomology
ring of the closed
oriented
surface X of genus 
g ≥ 0 has a basis as
a free Z-module of
the form: the
element 1 in degree
0, A1,...,Ag and B1,...,
Bg in degree 1, and
the class P of a
point in degree 2.
The product is
given
by: AiAj = BiBj = 0 for
all i and j, AiBj = 0
if i ≠ j,
and AiBi = P for all i.
[11]
 By graded-
commutativity, it
follows that BiAi =
−P.
 On any topological
space, graded-
commutativity of the
cohomology ring
implies that 2x2 = 0
for all odd-degree
cohomology
classes x. It follows
that for a
ring R containing
1/2, all odd-degree
elements of H*(X,R)
have square zero.
On the other hand,
odd-degree
elements need not
have square zero
if R is Z/2 or Z, as
one sees in the
example
of RP2 (with Z/2
coefficients)
or RP4 × RP2 (with Z 
coefficients).

The
diagonal[edit]
The cup product on
cohomology can be
viewed as coming
from the diagonal
map Δ: X → X × X, x 
↦ (x,x). Namely, for
any
spaces X and Y with
cohomology
classes u ∈ Hi(X,R)
and v ∈ Hj(Y,R), there
is an external
product (or cross
product) cohomology
class u × v ∈ Hi+j(X × 
Y,R). The cup product
of classes u ∈ Hi(X,R)
and v ∈ Hj(X,R) can
be defined as the
pullback of the
external product by
the diagonal:[12]
Alternatively, the
external product
can be defined in
terms of the cup
product. For
spaces X and Y,
write f: X × Y → X 
and g: X × Y → Y 
for the two
projections. Then
the external
product of
classes u ∈ Hi(X,
R) and v ∈ Hj(Y,R)
is:

Poincaré
duality[edit
]
Another
interpretation
of Poincaré
duality is that
the
cohomology
ring of a
closed
oriented
manifold is
self-dual in a
strong sense.
Namely,
let X be a
closed connec
ted oriented
manifold of
dimension n,
and let F be a
field.
Then Hn(X,F)
is isomorphic
to F, and the
product
is a perfect
pairing for
each
integer i.[13] 
In
particular,
the vector
spaces Hi(
X,F)
and Hn−i(X,
F) have
the same
(finite)
dimension.
Likewise,
the product
on integral
cohomolog
y
modulo tor
sion with
values
in Hn(X,Z)
≅ Z is a
perfect
pairing
over Z.

Charac
teristic
classes[
edit]
An
oriented
real vector
bundle E o
f
rank r over
a
topological
space X de
termines a
cohomolog
y class
on X,
the Euler
class χ(E)
∈ Hr(X,Z).
Informally,
the Euler
class is the
class of
the zero
set of a
general se
ction of E.
That
interpretati
on can be
made
more
explicit
when E is
a smooth
vector
bundle
over a
smooth
manifold X,
since then
a general
smooth
section
of X vanish
es on a
codimensi
on-r subm
anifold
of X.
There are
several
other types
of characte
ristic
classes for
vector
bundles
that take
values in
cohomolog
y,
including C
hern
classes, St
iefel–
Whitney
classes,
and Pontry
agin
classes.

Eilenbe
rg–
MacLa
ne
spaces[
edit]
For each
abelian
group A an
d natural
number j,
there is a
space K(A,
j)
whose jth
homotopy
group is
isomorphic
to A and
whose
other
homotopy
groups are
zero. Such
a space is
called
an Eilenbe
rg–
MacLane
space.
This space
has the
remarkable
property
that it is
a classifyi
ng
space for
cohomolog
y: there is
a natural
element u 
of Hj(K(A,j)
,A), and
every
cohomolog
y class of
degree j on
every
space X is
the
pullback
of u by
some
continuous
map X → 
K(A,j).
More
precisely,
pulling
back the
class u giv
es a
bijection
for
every
space 
X with
the
homoto
py type
of a
CW
comple
x.[14] Her
e [X,Y]
denote
s the
set of
homoto
py
classes
of
continu
ous
maps
from X 
to Y.
For
exampl
e, the
space 
K(Z,1)
(define
d up to
homoto
py
equival
ence)
can be
taken
to be
the
circle S
1
. So
the
descrip
tion
above
says
that
every
elemen
t
of H1(X,
Z) is
pulled
back
from
the
class u 
of a
point
on S1 b
y some
map X 
→ S1.
There
is a
related
descrip
tion of
the first
cohom
ology
with
coeffici
ents in
any
abelian
group 
A, say
for a
CW
comple
x X.
Namely
, H1(X,
A) is in
one-to-
one
corresp
ondenc
e with
the set
of
isomor
phism
classes
of
Galois 
coverin
g
spaces 
of X wit
h
group 
A, also
called 
princip
al A-
bundle
s over 
X.
For X c
onnect
ed, it
follows
that H1(
X,A) is
isomor
phic to
Hom(π1
X,A),
where
π1X is
the fun
dament
al
group o
f X. For
exampl
e, H1(X,
Z/2)
classifi
es the
double
coverin
g
spaces
of X,
with
the
elemen
t0
∈ H1(X,
Z/2)
corresp
onding
to the
trivial
double
coverin
g, the
disjoint
union
of two
copies
of X.

Cap
prod
uct[e
dit]
For any
topolog
ical
space 
X,
the cap
produc
t is a
bilinear
map
for
any
inte
ger
s i 
and 
j an
d
any
co
mm
utat
ive
ring 
R.
The
res
ulti
ng
ma
p
m
a
k
e
s

t
h
e

s
i
n
g
u
l
a
r
h
o
m
o
l
o
g
y

o
f
 
X
 
i
n
t
o

m
o
d
u
l
e

o
v
e
r

t
h
e

s
i
n
g
u
l
a
r

c
o
h
o
m
o
l
o
g
y

r
i
n
g

o
f
 
X
.
F
o
r
 
i
 
=
 
j
,

t
h
e

c
a
p

p
r
o
d
u
c
t

g
i
v
e
s
t
h
e

n
a
t
u
r
a
l
h
o
m
o
m
o
r
p
h
i
s
m
whi
ch
is
an
iso
mo
rph
ism
for 
R a
fiel
d.
For
ex
am
ple
,
let 
X b
e
an
ori
ent
ed
ma
nif
old
,
not
ne
ces
sar
ily
co
mp
act
.
Th
en
a
clo
se
d
ori
ent
ed
co
di
me
nsi
on-
i su
bm
ani
fol

Y o
f X 
(no
t
ne
ces
sar
ily
co
mp
act
)
det
er
mi
ne
s
an
ele
me
nt
of 
H i(
X,
R),
an
da
co
mp
act
ori
ent
ed 
j-
di
me
nsi
on
al
su
bm
ani
fol
d Z 
of 
X d
ete
rmi
ne
s
an
ele
me
nt
of 
H j(
X,
R).
Th
e
ca
p
pro
du
ct
[Y]

[Z]
∈ 
Hj−i(
X,
R)
ca
n
be
co
mp
ute
d
by
per
tur
bin

Y a
nd 
Z t
o
ma
ke
the
m
int
ers
ect
tra
nsv
ers
ely
an
d
the
n
taki
ng
the
cla
ss
of
the
ir
int
ers
ecti
on,
whi
ch
is a
co
mp
act
ori
ent
ed
su
bm
ani
fol
d
of
di
me
nsi
on 
j − 
i.
A
clo
se
d
ori
ent
ed
ma
nif
old 
X o
f
di
me
nsi
on 
n h
as
a f
un
da
me
nta
l
cla
ss [
X]
in 
Hn(
X,
R).
Th
e
Poi
nc
aré
du
alit
y
iso
mo
rph
ism
is
define
d by
cap
produc
t with
the
funda
mental
class
of X.

Hist
ory,
to
the
birt
h of
sing
ular
coh
omo
logy[
edit]
Althou
gh
cohom
ology
is
funda
mental
to
moder
n
algebr
aic
topolo
gy, its
import
ance
was
not
seen
for
some
40
years
after
the
develo
pment
of
homol
ogy.
The
conce
pt
of dual
cell
structu
re,
which 
Henri
Poinca
ré use
d in
his
proof
of his
Poinca

duality
theore
m,
contai
ned
the
germ
of the
idea of
cohom
ology,
but
this
was
not
seen
until
later.
There
were
variou
s
precur
sors to
cohom
ology.
[15]
 In
the
mid-
1920s, 
J. W.
Alexan
der an
d Solo
mon
Lefsch
etz fou
nded
the int
ersecti
on
theory 
of
cycles
on
manifo
lds.
On a
closed
oriente
d n-
dimen
sional
manifo
ld M,
an i-
cycle
and
a j-
cycle
with
nonem
pty
interse
ction
will, if
in gen
eral
positio
n,
have
interse
ction
an
(i + j − 
n)-
cycle.
This
leads
to a
multipli
cation
of
homol
ogy
classe
s
which in
retrospect
can be
identified
with the
cup
product on
the
cohomolo
gy of M.
Alexander
had by
1930
defined a
first notion
of a
cochain,
by
thinking of
an i-
cochain
on a
space X a
sa
function
on small
neighborh
oods of
the
diagonal
in Xi+1.
In
1931, Geo
rges de
Rham rela
ted
homology
and
differential
forms,
proving de
Rham's
theorem.
This result
can be
stated
more
simply in
terms of
cohomolo
gy.
In
1934, Lev
Pontryagi
n proved
the Pontry
agin
duality the
orem; a
result
on topolog
ical
groups.
This (in
rather
special
cases)
provided
an
interpretati
on of
Poincaré
duality
and Alexa
nder
duality in
terms of
group cha
racters.
At a 1935
conferenc
e
in Moscow
, Andrey
Kolmogor
ov and
Alexander
both
introduced
cohomolo
gy and
tried to
construct
a
cohomolo
gy product
structure.
In
1936, Nor
man
Steenrod 
constructe
d Čech
cohomolo
gy by
dualizing
Čech
homology.
From
1936 to
1938, Has
sler
Whitney a
nd Eduard
Čech dev
eloped the
cup
product
(making
cohomolo
gy into a
graded
ring) and
cap
product,
and
realized
that
Poincaré
duality
can be
stated in
terms of
the cap
product.
Their
theory
was still
limited to
finite cell
complexe
s.
In
1944, Sa
muel
Eilenberg 
overcame
the
technical
limitations,
and gave
the
modern
definition
of singular
homology
and
cohomolo
gy.
In 1945,
Eilenberg
and
Steenrod
stated
the axiom
s defining
a
homology
or
cohomolo
gy theory,
discussed
below. In
their 1952
book, Fou
ndations
of
Algebraic
Topology,
they
proved
that the
existing
homology
and
cohomolo
gy
theories
did indeed
satisfy
their
axioms.
In
1946, Jea
n
Leray defi
ned sheaf
cohomolo
gy.
In
1948 Edwi
n Spanier,
building
on work of
Alexander
and
Kolmogor
ov,
developed 
Alexander
–Spanier
cohomolo
gy.

Sheaf
cohom
ology[e
dit]
Sheaf
cohomol
ogy is a
rich
generaliza
tion of
singular
cohomolo
gy,
allowing
more
general
"coefficien
ts" than
simply an
abelian
group. For
every she
af of
abelian
groups E 
on a
topologica
l space X,
one has
cohomolo
gy
groups Hi(
X,E) for
integers i.
In
particular,
in the
case of
the consta
nt
sheaf on 
X associat
ed to an
abelian
group A,
the
resulting
groups Hi(
X,A)
coincide
with
singular
cohomolo
gy for X a
manifold
or CW
complex
(though
not for
arbitrary
spaces X).
Starting in
the 1950s,
sheaf
cohomolo
gy has
become a
central
part
of algebrai
c
geometry 
and compl
ex
analysis,
partly
because
of the
importanc
e of the
sheaf
of regular
functions 
or the
sheaf
of holomor
phic
functions.
Grothendi
eck elega
ntly
defined
and
characteri
zed sheaf
cohomolo
gy in the
language
of homolo
gical
algebra.
The
essential
point is to
fix the
space X a
nd think of
sheaf
cohomolo
gy as a
functor
from
the abelia
n
category o
f sheaves
on X to
abelian
groups.
Start with
the functor
taking a
sheaf E o
n X to its
abelian
group of
global
sections
over X, E(
X). This
functor
is left
exact, but
not
necessaril
y right
exact.
Grothendi
eck
defined
sheaf
cohomolo
gy groups
to be the
right deriv
ed
functors of
the left
exact
functor E 
↦ E(X).[16]
That
definition
suggests
various
generaliza
tions. For
example,
one can
define the
cohomolo
gy of a
topologica
l
space X w
ith
coefficient
s in any
complex
of
sheaves,
earlier
called hyp
ercohomol
ogy (but
usually
now just
"cohomolo
gy"). From
that point
of view,
sheaf
cohomolo
gy
becomes
a
sequence
of functors
from
the derive
d
category o
f sheaves
on X to
abelian
groups.
In a broad
sense of
the word,
"cohomolo
gy" is
often used
for the
right
derived
functors of
a left
exact
functor on
an abelian
category,
while
"homology
" is used
for the left
derived
functors of
a right
exact
functor.
For
example,
for a
ring R,
the Tor
groups To
riR(M,N)
form a
"homology
theory" in
each
variable,
the left
derived
functors of
the tensor
product M
⊗RN of R-
modules.
Likewise,
the Ext
groups Ex
tiR(M,N)
can be
viewed as
a
"cohomolo
gy theory"
in each
variable,
the right
derived
functors of
the Hom
functor
HomR(M,N
).
Sheaf
cohomolo
gy can be
identified
with a
type of Ext
group.
Namely,
for a
sheaf E o
na
topologica
l
space X, 
Hi(X,E) is
isomorphi
c to
Exti(ZX, E),
where ZX d
enotes the
constant
sheaf
associate
d to the
integers Z,
and Ext is
taken in
the
abelian
category
of
sheaves
on X.

Coho
molog
y of
varieti
es[edit]
There are
numerous
machines
built for
computing
the
cohomolo
gy of
algebraic
varieties.
The
simplest
case
being the
determinat
ion of
cohomolo
gy for
smooth
projective
varieties
over a
field of
characteri
stic .
Tools from
Hodge
theory,
called Hod
ge
structures 
help give
computati
ons of
cohomolo
gy of
these
types of
varieties
(with the
addition of
more
refined
informatio
n). In the
simplest
case the
cohomolo
gy of a
smooth
hypersurfa
ce in  can
be
determine
d from the
degree of
the
polynomia
l alone.
When
considerin
g varieties
over a
finite field,
or a field
of
characteri
stic , more
powerful
tools are
required
because
the
classical
definitions
of
homology/
cohomolo
gy break
down.
This is
because
varieties
over finite
fields will
only be a
finite set
of points.
Grothendi
eck came
up with
the idea
for a
Grothendi
eck
topology
and used
sheaf
cohomolo
gy over
the etale
topology
to define
the
cohomolo
gy theory
for
varieties
over a
finite field.
Using the
étale
topology
for a
variety
over a
field of
characteri
stic  one
can
construct -
adic
cohomolo
gy for .
This is
defined as
If we have a
scheme of
finite type
then there is an
equality of
dimensions for
the Betti
cohomology
of  and the -adic
cohomology
of  whenever the
variety is smooth
over both fields.
In addition to
these
cohomology
theories there ar
other cohomolog
theories
called Weil
cohomology
theories which
behave similarly
to singular
cohomology.
There is a
conjectured
theory of motives
which underlie a
of the Weil
cohomology
theories.
Another useful
computational
tool is the blowup
sequence. Given
a
codimension 
 
subscheme  ther
is a Cartesian
square
From this there i
associated long
exact sequence
If the subvariety 
smooth, then the
connecting morp
are all trivial, hen

Axioms and
generalized
cohomolog
theories[edit
See also: List of c
theories

There are variou


define cohomolo
topological space
singular cohomo
cohomology, Ale
Spanier
cohomology or s
cohomology). (H
cohomology is co
only with coeffici
constant sheaf.)
theories give diff
answers for som
but there is a larg
spaces on which
agree. This is mo
understood axiom
there is a list of p
known as the Eil
Steenrod axioms
two construction
those properties
at least on all CW
complexes.[17] Th
versions of the a
homology theory
for a cohomology
Some theories c
viewed as tools f
computing singu
cohomology for s
topological space
as simplicial
cohomology for s
complexes, cellu
cohomology for C
complexes, and 
cohomology for s
manifolds.
One of the Eilenb
Steenrod axioms
cohomology theo
the dimension a
if P is a single po
then Hi(P) = 0 fo
Around 1960, Ge
Whitehead obse
is fruitful to omit
dimension axiom
completely: this g
notion of a gener
homology theory
generalized coho
theory, defined b
are generalized c
theories such as
complex cobordi
rich information a
topological space
directly accessib
singular cohomo
context, singular
cohomology is o
"ordinary cohom
By definition, a g
homology theor
sequence of func
integers i) from
the category of C
pairs (X, A) (so X
complex and A is
subcomplex) to t
of abelian groups
with a natural
transformation ∂i
→ hi−1(A) called
the boundary
homomorphism
) is a shorthand f
The axioms are:
1. Homotopy: If 
then the induc
homomorphism
are the same.
2. Exactness: Ea
induces a long
in homology, v
inclusions f: A 
→ (X,A):

3. Excision: If X 
subcomplexes
the inclusion f:
(X,B) induces

for every i.

4. Additivity: If (
disjoint union o
(Xα,Aα), then th
(Xα,Aα) → (X,A
isomorphism f
sum:

for every i.
The axioms for a
generalized coho
theory are obtain
reversing the arr
roughly speaking
detail, a general
cohomology the
sequence of con
functors hi (for in
from the categor
pairs to the categ
abelian groups, t
with a natural
transformation d
→ hi+1(X,A) called
the boundary
homomorphism
A) for hi(A,∅)). Th
are:

1. Homotopy: H
induce the sam
homomorphism
cohomology.
2. Exactness: Ea
induces a long
in cohomology
inclusions f: A 
→ (X,A):

3. Excision: If X 
subcomplexes
the inclusion f:
(X,B) induces

for every i.

4. Additivity: If (
disjoint union o
(Xα,Aα), then th
(Xα,Aα) → (X,A
isomorphism t
group:

for every i.
A spectrum dete
a generalized ho
theory and a gen
cohomology theo
fundamental resu
Brown, Whitehea
and Adams says
generalized hom
theory comes fro
spectrum, and lik
every generalize
cohomology theo
from a spectrum
generalizes the
representability o
cohomology by E
MacLane spaces
A subtle point is
functor from the
homotopy catego
homotopy catego
spectra) to gene
homology theorie
pairs is not an eq
although it gives
on isomorphism
there are nonzer
the stable homot
category (called 
maps) that induc
map between ho
theories on CW-
Likewise, the fun
the stable homot
category to gene
cohomology theo
CW-pairs is not a
equivalence.[19] It
homotopy catego
these other categ
has good proper
being triangulate
If one prefers ho
cohomology theo
defined on all top
spaces rather tha
complexes, one
approach is to in
axiom that every
homotopy
equivalence indu
isomorphism on
or cohomology. (
for singular homo
singular cohomo
not for sheaf coh
for example.) Sin
space admits a w
homotopy equiva
a CW complex, t
reduces homolog
cohomology theo
spaces to the co
theory on CW co
Some examples
generalized coho
theories are:
 Stable cohomo
groups  The co
homology theo
more often: sta
homotopy grou
 Various differe
of cobordism g
on studying a
considering all
to manifolds: u
cobordism  ori
cobordism  co
cobordism  an
Complex cobo
turned out to b
powerful in ho
theory. It is clo
to formal group
theorem of Da
 Various differe
topological K-t
on studying a
considering all
bundles over i
periodic K-theo
connective K-
theory),  (comp
K-theory),  (co
connective K-t
so on.
 Brown–Peters
cohomology, M
theory, Morava
and other theo
from complex
 Various flavors
cohomology.
Many of these th
richer information
ordinary cohomo
are harder to com
A cohomology th
said to
be multiplicativ
structure of a gra
each space X. In
language of spec
are several more
notions of a ring
such as an E∞ rin
where the produ
commutative and
associative in a s
sense.

Other coho
theories[edit
Cohomology the
broader sense (i
other algebraic o
structures, rathe
topological space

 Algebraic K-th
 André–Quillen
 BRST cohomo
 Coherent shea
cohomology
 Crystalline coh
 Cyclic cohomo
 Deligne cohom
 Equivariant co
 Étale cohomol
 Ext groups
 Flat cohomolo
 Floer homolog
 Galois cohomo
 Group cohomo
 Hochschild co
 Intersection co
 Khovanov hom
 Lie algebra co
 Local cohomo
 Motivic cohom
 Non-abelian co
 Quantum coho

See also[edit
 complex-orien
cohomology th

Notes[edit]
1. ^ Hatcher (2001
2. ^ Hatcher (2001
Dold (1972), Pro
Corollary VIII.3.4
3. ^ Dold (1972), P
and V.4.11.
4. ^ Hatcher (2001
5. ^ Thom (1954), p
6. ^ Thom (1954), T
7. ^ Hatcher (2001
8. ^ Hatcher (2001
9. ^ Jump up to:a b Hatc
Theorem 3.19.
10. ^ Hatcher (2001
11. ^ Hatcher (2001
12. ^ Hatcher (2001
13. ^ Hatcher (2001
14. ^ May (1999), p.
15. ^ Dieudonné (19
16. ^ Hartshorne (19
17. ^ May (1999), p.
18. ^ Switzer (1975)
Corollary 14.36;
and p. 331.
19. ^ "Are spectra re
cohomology
theories?".  Math
20. ^ Switzer (1975)

References
 Dieudonné,
Jean (1989), H
Algebraic and
Topology, Birk
N 0-8176-338
5842
 Dold,
Albrecht (1972
on Algebraic
Topology, Spr
ISBN 978-3-54
MR 0415602
 Eilenberg,
Samuel; Steen
Norman (1952
ns of Algebraic
Topology, Prin
University
Press, ISBN 9
36, MR 00508
 Hartshorne,
Robin (1977), 
Geometry, Gra
in Mathematic
York, Heidelbe
Verlag, ISBN 
9, MR 046315
 Hatcher,
Allen (2001), A
Topology, Cam
University Pre
521-79540-0, 
 Hazewinkel, M
(2001)
[1994], "Cohom
yclopedia of M
Springer Scien
Media B.V. / K
Academic
Publishers, IS
55608-010-4.
 May, J. Peter 
Concise Cours
Algebraic
Topology  (PDF
of Chicago Pre
226-51182-0, 
 Switzer, Robe
(1975), Algebr
— Homology a
Homotopy, Sp
Verlag, ISBN 
3, MR 038583
 Thom,
René (1954), 
propriétés glob
variétés
différentiables
rii Mathematic
Helvetici, 28: 1
86, doi:10.100
23, MR 00618

hide

 v

 t

 e

Вам также может понравиться