Normalized 2-cocycle for trivial group action: Difference between revisions

From Groupprops
 
(29 intermediate revisions by the same user not shown)
Line 1: Line 1:
==Definition==
==Definition==


Suppose <math>G</math> is a [[group]] and <math>A</math> is an [[abelian group]]. A '''normalized 2-cocycle for trivial group action''' is a function <math>f: G \times G \to A</math> that satisfies both these conditions:
===Equivalent definitions in tabular format===


* <math>f</math> is a [[defining ingredient::2-cocycle for trivial group action]]: <math>f(g,hk) + f(h,k) = f(gh,k) + f(g,h) \ \forall \ g,h,k, \in G</math>
Suppose <math>G</math> is a [[group]] and <math>A</math> is an [[abelian group]]. We denote by <math>1</math> the identity element of <math>G</math> and by <math>0</math> the identity element of <math>A</math>; the group operation for <math>G</math> is denoted multiplicatively and the group operation for <math>A</math> is denoted additively. A '''normalized 2-cocycle for trivial group action''' is a function <math>f: G \times G \to A</math> that satisfies the following equivalent conditions:
* <math>f</math> evaluates to the identity element of <math>A</math> (written as 0 since <math>A</math> is abelian) if either of its inputs is the identity element of <math>G</math>: <math>f(g, 1) = f(1, g) = 0 \ \forall \ g \in G</math>


The set of normalized 2-cocycles for trivial group action form a subgroup (under pointwise addition) of the group of 2-cocycles for trivial group action. We will denote this subgroup as <math>Z^2_n(G, A)</math>, though this is not standard notation.


Although this is a proper subgroup whenever <math>A</math> is nontrivial (and in fact the quotient group <math>Z^2(G, A) / Z^2_n(G, A)</math> is isomorphic to <math>A</math>), it still maps surjectively to the [[second cohomology group for trivial group action]], as explained in the Importance section.
{| class="sortable" border="1"
! No. !! Shorthand !! Full statement
|-
| 1 || 2-cocycle for trivial group action that is zero along the axes || <math>f</math> is a [[defining ingredient::2-cocycle for trivial group action]]: <math>f(g,hk) + f(h,k) = f(gh,k) + f(g,h) \ \forall \ g,h,k, \in G</math> '''and''' <math>f(g,1) = f(1,g) = 0 \ \forall \ g \in G</math>.
|-
| 2 || 2-cocycle for trivial group action that is zero at origin || <math>f</math> is a 2-cocycle for trivial group action: <math>f(g,hk) + f(h,k) = f(gh,k) + f(g,h) \ \forall \ g,h,k, \in G</math> '''and''' <math>f(1,1) = 0</math>.
|-
| 3 || System of coset representatives where identity is represented by identity || <math>f</math> corresponds to a system of coset representatives for a central extension <math>E</math> (corresponding to an element of <math>H^2(G,A)</math>, the [[second cohomology group for trivial group action]]) where the coset representative of the identity element is the identity element.
|}
{{tabular definition format}}


==Importance==
The set of normalized 2-cocycles for trivial group action forms a subgroup (under pointwise addition) of the group of 2-cocycles for trivial group action. We will denote this subgroup as <math>Z^2_n(G, A)</math>, though this is not standard notation. For the relationship with 2-coboundaries, constant functions, and second cohomology, see [[#Importance]].
 
===Equivalence of definitions===


===Normalizing a 2-cocycle===
The equivalence of definitions (1) and (2) follows from [[2-cocycle for trivial group action is constant on axes]], which states that <math>f(g,1) = f(1,g) = f(1,1) \ \forall \ g \in G</math>.


====Background fact====
The equivalence with definition (3) is discussed in more detail at [[#Relation with central extensions]].


One important background fact relevant to understanding normalization is that for any 2-cocycle <math>f: G \times G \to A</math>, <math>f(x, 1) = f(1, x) = f(1, 1)</math> for all <math>x \in G</math>
==Importance==


* The fact that <math>f(x, 1) = f(1, 1)</math> can be proved by setting <math>g = x, h = 1, k = 1</math> in the identity <math>f(g, hk) + f(h, k) = f(gh, k) + f(g, h)</math> and getting <math>f(x, 1) + f(1, 1) = f(x, 1) + f(x, 1)</math>. One <math>f(x, 1)</math> can be cancelled from both sides, leaving <math>f(1, 1) = f(x, 1)</math>.
===Normalizing a 2-cocycle (finding a normalized 2-cocycle that differs from it via a 2-coboundary)===
* The fact that <math>f(1, 1) = f(1, x)</math> can be proved by setting <math>g = 1, h = 1, k = x</math> in the identity <math>f(g, hk) + f(h, k) = f(gh, k) + f(g, h)</math> and getting <math>f(1, x) + f(1, x) = f(1, x) + f(1, 1)</math>. One <math>f(1, x)</math> can be cancelled from both sides, leaving <math>f(1, 1) = f(1, x)</math>.


====Translation normalization====
====Translation normalization====
Line 25: Line 33:
The simplest normalization rule is a translation rule: given a 2-cocycle for trivial group action <math>f: G \times G \to A</math>, we can obtain a normalized 2-cocycle <math>f_0:G \times G \to A</math> defined as follows:
The simplest normalization rule is a translation rule: given a 2-cocycle for trivial group action <math>f: G \times G \to A</math>, we can obtain a normalized 2-cocycle <math>f_0:G \times G \to A</math> defined as follows:


<math>f_0(g, h) = f(g, h) - f(1, 1)</math>
<math>f_0(g,h) = f(g,h) - f(1,1)</math>


<math>f_0</math> satisfies the identity for a 2-cocycle because it's just taking the original identity and subtracting <math>f(1,1)</math> twice from both sides.
<math>f_0</math> satisfies the identity for a 2-cocycle because it's just taking the original identity and subtracting <math>f(1,1)</math> twice from both sides.


<math>f_0</math> is also normalized, because <math>f_0(g, 1) = f(g, 1) - f(1, 1)</math>, which, based on the background fact, works out to zero.
<math>f_0</math> is also normalized, because <math>f_0(g,1) = f(g,1) - f(1,1)</math> and <math>f_0(1,g) = f(1,g) - f(1,1)</math>, which are both zero based on [[2-cocycle for trivial group action is constant on axes]].
 
Moreover, the difference of <math>f</math> and <math>f_0</math> is a constant function <math>G \times G \to A</math> with constant value <math>f(1,1)</math>, which is therefore also a 2-coboundary (of the constant function <math>G \to A</math> taking value <math>f(1,1)</math>. Thus, <math>f_0</math> belongs to the same cohomology class as <math>f</math>, i.e., they project to the same element of the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math>.


====Alternative normalization====
====Alternative normalization====
Line 35: Line 45:
Given a 2-cocycle for trivial group action <math>f: G \times G \to A</math>, we can obtain a normalized 2-cocycle <math>f_1:G \times G \to A</math> defined as follows:
Given a 2-cocycle for trivial group action <math>f: G \times G \to A</math>, we can obtain a normalized 2-cocycle <math>f_1:G \times G \to A</math> defined as follows:


* <math>f_1(g, h) = f(g, h)</math> if <math>g, h, gh</math> are all non-identity elements
* <math>f_1(g,h) = f(g,h)</math> if <math>g, h, gh</math> are all non-identity elements
* <math>f_1(g, 1) = f_1(1, g) = 0</math> for all <math>g</math>
* <math>f_1(g,1) = f_1(1,g) = 0</math> for all <math>g</math>
* <math>f_1(g, g^{-1}) = f(g, g^{-1}) + f(1, 1)</math> for all <math>g</math> (note that it follows from the definition of 2-cocycle that <math>f(g, 1) = f(1, 1)</math> for all <math>g</math>, so this can also be written as <math>f_1(g, g^{-1}) = f(g, g^{-1}) + f(g, 1)</math>)
* <math>f_1(g,g^{-1}) = f(g,g^{-1}) + f(1, 1)</math> for all <math>g</math> (note that it follows from the definition of 2-cocycle that <math>f(g, 1) = f(1, 1)</math> for all <math>g</math>, so this can also be written as <math>f_1(g,g^{-1}) = f(g,g^{-1}) + f(g,1)</math>)


<math>f_1</math> is different from <math>f_0</math>. As the later discussion will clarify, they differ by a [[2-coboundary for trivial group action]].
<math>f_1</math> is different from <math>f_0</math>. As the later discussion will clarify, they differ by a [[2-coboundary for trivial group action]].


===Relation with treatment as central extensions===
For most real-world purposes, we go with <math>f_0</math>. However, <math>f_1</math> could be useful if we want to preserve as many of the values of <math>f</math> as possible.
 
===The group of 2-cocycles is a direct sum of the subgroup of normalized 2-cocycles and the base group (realized as constant functions)===
 
We have:
 
<math>Z^2(G,A) = Z^2_n(G,A) \oplus A</math>
 
This is a canonical split as an internal direct sum. The terms are as follows:
 
* <math>Z^2(G,A)</math> is the group of all 2-cocycles for trivial group action.
* <math>Z^2_n(G,A)</math> is the subgroup comprising normalized 2-cocycles for trivial group action.
* <math>A</math> is identified with the subgroup comprising constant functions <math>G \times G \to A</math>. Specifically, an element <math>a \in A</math> corresponds to the constant function <math>c_a</math> sending everything in <math>G \times G</math> to <math>a</math>.
 
To show that this is an internal direct sum, we need to verify the following:
 
* <math>Z^2_n(G,A)</math> is a subgroup of <math>Z^2(G,A)</math>: This is obvious from the definition.
* <math>A</math> (identified as constant functions) is a subgroup of <math>Z^2(G,A)</math>: This is also obvious from the definition of 2-cocycle.
* <math>Z^2_n(G,A) \cap A = 0</math>: For <math>f</math> a normalized 2-cocycle, <math>f(1,1) = 0</math>. If <math>f</math> is also constant, this forces it to be zero everywhere.
* <math>Z^2_n(G,A) + A  = Z^2(G,A)</math>, i.e., every 2-cocycle can be expressed as the sum of a normalized 2-cocycle and a constant function: We already demonstrated this in the preceding section in the form of translation normalization.
 
Some implications of this:
 
* <math>Z^2(G,A) / Z^2_n(G,A) \cong A</math>
* <math>Z^2(G,A) / A \cong Z^2_n(G,A)</math> (where the group <math>A</math> in the denominator is realized as constant functions)
* <math>0 \to Z^2_n(G,A) \to Z^2(G,A) \to A \to 0</math> is a canonically split short exact sequence.
* <math>0 \to A \to Z^2(G,A) \to Z^2_n(G,A) \to 0</math> is a canonically split short exact sequence.
 
===Putting together 2-coboundaries, normalized 2-cocycles, and cocycles===
 
Let's articulate the notation:
 
* <math>Z^2(G, A)</math> is a group of [[2-cocycle for trivial group action|2-cocycles for trivial group action]].
* <math>Z^2_n(G, A)</math> is the subgroup comprising normalized 2-cocycles.
* <math>B^2(G, A)</math> is the subgroup comprising 2-coboundaries.
* <math>B^2_n(G, A)</math> is the subgroup comprising normalized 2-coboundaries; it is the intersection <math>Z^2_n(G,A) \cap B^2(G,A)</math>.
* <math>A</math> is identified with constant functions <math>G \times G \to A</math>.
* <math>H^2(G, A)</math> is the quotient <math>Z^2(G, A) / B^2(G, A)</math> and is called the [[second cohomology group for trivial group action]].
 
We have two short exact sequences, the first of which splits canonically and was discussed in the preceding section:
 
<math>0 \to A \to Z^2(G, A) \to Z^2_n(G, A) \to 0</math>
 
<math>0 \to B^2(G, A) \to Z^2(G, A) \to H^2(G, A) \to 0</math>


Let <math>E</math> be a group with a [[central subgroup]] isomorphic to (and explicitly identified with) <math>A</math>, and a quotient isomorphic to (and explicitly identified with) <math>G</math>, such that the induced action of the quotient on the subgroup (in the sense of action by conjugation, see [[quotient group acts on abelian normal subgroup]]). Let <math>S</math> be a system of coset representatives for <math>G</math> in <math>E</math> with <math>s: G \to S</math> being the representation map. Then, define <math>f: G \times G \to A</math> such that
The new fact of interest is that all constant functions are also 2-coboundaries! That's because the constant function <math>G \times G \to A</math> taking value <math>a \in A</math> can be realized as the coboundary of the constant function <math>G \to A</math> taking value <math>a</math>. This means that <math>A</math> maps injectively into <math>B^2(G,A)</math>. This allows us to make a downward arrow from <math>A</math> to <math>B^2(G, A)</math> that is injective. We can also make a downward arrow between the middle groups <math>Z^2(G,A)</math> that is the identity map. This in turn allows us to make a downward map from <math>Z^2_n(G, A)</math> to <math>H^2(G, A)</math>, and shows that this map must be surjective. In other words, every cohomology class can be represented by a normalized 2-cocycle. The kernel of this map is <math>B^2_n(G, A)</math>, so we obtain this short exact sequence:


<math>s(gh) = f(g,h)s(g)s(h)</math>
<math>0 \to B^2_n(G, A) \to Z^2_n(G, A) \to H^2(G, A) \to 0</math>


In other words, <math>f</math> measures the extent to which the collection of coset representatives fails to be closed under multiplication.
We also have a canonical internal direct sum splitting:


Such an <math>f</math> is a [[2-cocycle for trivial group action]] of <math>G</math> on <math>A</math>. The different possible 2-cocycles for a given central extension form a single coset of the subgroup of 2-coboundaries, and hence a single element of the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math>.
<math>B^2(G, A) = B^2_n(G, A) \oplus A</math>


If we further restrict <math>s</math> so that <math>s(1)</math> is the identity element of <math>A</math>, i.e., the coset representative for the base subgroup is its identity element, then the 2-cocycle we obtain this way is a normalized 2-cocycle.
This uses the same logic as the splitting for <math>Z^2(G, A)</math> from the preceding section, and the fact that all the constant functions are in fact 2-coboundaries.


Since it's always possible to choose <math>s</math> so that <math>s(1)</math> is the identity element of <math>A</math>, every central extension can be represented by a normalized 2-cocycle. In other words, every element of the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math> can be represented by a normalized 2-cocycle. Another way to frame this is that <math>Z^2(G, A)</math> (the group of all 2-cocycles for trivial group action) is generated by <math>B^2(G, A)</math> (the subgroup comprising 2-coboundaries) and the subgroup comprising normalized 2-cocycles.
We can put all this information together in the commutative diagram below, whose rows are short exact sequences and the last column is also a short exact sequence; the first two rows also have canonical splittings:


We can explain the two normalizations discussed above as follows:
<math>\begin{pmatrix}
      &            &    &            &    & 0          &    &  \\
      &            &    &            &    & \downarrow  &    &  \\
0 \to & A          & \to & Z^2_n(G, A) & \to & B^2_n(G, A) & \to & 0 \\
      & \downarrow &    & \downarrow  &    & \downarrow  &    &  \\
0 \to & A          & \to & Z^2(G, A)  & \to & Z^2_n(G, A) & \to & 0 \\
      & \downarrow &    & \downarrow  &    & \downarrow  &    &  \\
0 \to & B^2(G, A)  & \to & Z^2(G, A)  &\to  & H^2(G, A)  & \to & 0 \\
      &            &    &            &    & \downarrow  &    &  \\
      &            &    &            &    & 0          &    &  \\
\end{pmatrix}
</math>


* The translation normalization corresponds to taking the original system of coset representatives and multiplying all of them by <math>s(1)^{-1}</math> (left and right doesn't matter because <math>s(1)</math> is central).
===Relation with treatment as central extensions===
* The alternative normalization corresponds to keeping all coset representatives intact except <math>s(1)</math>, which is replaced by the identity element.


===Quotient group of 2-cocycles by normalized 2-cocycles is the base group; in fact, a short exact sequence===
This subsection looks at 2-cocycles in terms of their interpretation in terms of central extensions. It leads to a restatement / rediscovery / different perspective on some of the results of the preceding subsections.


We can actually say something stronger. Namely, there is a short exact sequence that splits canonically:
Let <math>E</math> be a group with a [[central subgroup]] isomorphic to (and explicitly identified with) <math>A</math>, and a quotient isomorphic to (and explicitly identified with) <math>G</math>, such that the induced action of the quotient on the subgroup (in the sense of action by conjugation, see [[quotient group acts on abelian normal subgroup]]). Let <math>S</math> be a system of coset representatives for <math>G</math> in <math>E</math> with <math>s: G \to S</math> being the representation map. Then, define <math>f: G \times G \to A</math> such that


<math>0 \to Z^2_n(G, A) \to Z^2(G, A) \to A \to 0</math>
<math>s(gh) = f(g,h)s(g)s(h)</math>


The injection from <math>Z^2_n(G, A)</math> to <math>Z^2(G, A)</math> is just treating a normalized 2-cocycle as a 2-cocycle. The surjection <math>Z^2(G, A) \to A</math> maps <math>f</math> to <math>f(1, 1)</math>. This is a short exact sequence for the following reasons:
In other words, <math>f</math> measures the extent to which the collection of coset representatives fails to be closed under multiplication.


* Injectivity of <math>Z^2_n(G, A) \to Z^2(G, A)</math>: This is obvious
Such an <math>f</math> is a [[2-cocycle for trivial group action]] of <math>G</math> on <math>A</math>. The different possible 2-cocycles for a given central extension form a single coset of the subgroup of 2-coboundaries, and hence a single element of the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math>.
* Composite <math>Z^2_n(G, A) \to Z^2(G, A) \to A</math> is zero: This follows from the definition of normalized
* Middle exactness: If <math>f(1, 1) = 0</math>, then by the background fact above, <math>f(g, 1) = f(1, g) = 0</math> as well, which is precisely the definition of normalized 2-cocycle.
* Surjectivity and canonical splitting of <math>Z^2(G, A) \to A</math>: For any element <math>a \in A</math>, the constant function mapping everything to <math>A</math> is a 2-cocycle (in fact, as we will see shortly, it is a 2-coboundary) so every element of <math>A</math> is obtained this way.


===Putting together 2-coboundaries, normalized 2-cocycles, and cocycles===
The key insight is as follows:


Let's articulate the notation:
'''The 2-cocycle is a normalized 2-cocycle if and only if <math>s(1)</math> (which needs to be an element of <math>A</math> inside <math>E</math>) is the identity element of <math>E</math> (and hence also of <math>A</math> as a subgroup of <math>E</math>).'''


* <math>Z^2(G, A)</math> is a group of [[2-cocycle for trivial group action|2-cocycles for trivial group action]]
Since it's always possible to choose <math>s</math> so that <math>s(1)</math> is the identity element of <math>A</math>, every central extension can be represented by a normalized 2-cocycle. In other words, every element of the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math> can be represented by a normalized 2-cocycle. This is just a restatement of our earlier observation that <math>Z^2_n(G,A)</math> maps surjectively to <math>H^2(G,A)</math>.
* <math>Z^2_n(G, A)</math> is the subgroup comprising normalized 2-cocycles
* <math>B^2(G, A)</math> is the subgroup comprising 2-coboundaries
* <math>H^2(G, A)</math> is the quotient <math>Z^2(G, A) / B^2(G, A)</math> and is called the [[second cohomology group for trivial group action]]


We have two short exact sequences:
We can explain the two normalizations discussed above as follows:


<math>0 \to Z^2_n(G, A) \to Z^2(G, A) \to A \to 0</math>
* The translation normalization corresponds to taking the original system of coset representatives and multiplying all of them by <math>s(1)^{-1}</math> (left and right doesn't matter because <math>s(1)</math> is central).
* The alternative normalization corresponds to keeping all coset representatives intact except <math>s(1)</math>, which is replaced by the identity element.


<math>0 \to B^2(G, A) \to Z^2(G, A) \to H^2(G, A) \to 0</math>
==Relation with other properties==


The first of these splits canonically as well! So it can be reversed, and we can write:
===Stronger properties===


<math>0 \to A \to Z^2(G, A) \to Z^2_n(G, A) \to 0</math>
{| class="sortable" border="1"
! Property !! Meaning !! Proof of implication !! Proof of strictness !! Capture of difference !! Intermediate notions
|-
| [[Weaker than::cyclicity-preserving 2-cocycle for trivial group action]] || A 2-cocycle <math>f</math> such that <math>f(g,h) = 0</math> whenever <math>\langle g,h \rangle</math> is cyclic || || || || {{intermediate notions short|normalized 2-cocycle for trivial group action|cyclicity-preserving 2-cocycle for trivial group action}}
|-
| [[Weaker than::IIP 2-cocycle for trivial group action]] || A 2-cocycle <math>f</math> such that <math>f(g, g^{-1}) = 0</math> for all <math>g</math> || || || || {{intermediate notions short|normalized 2-cocycle for trivial group action|IIP 2-cocycle for trivial group action}}
|-
| [[Weaker than::normalized 2-coboundary for trivial group action]] || A 2-coboundary <math>f</math> such that <math>f(g, 1) = f(1, g) = 0</math> for all <math>g</math> || || || The quotient group is canonically isomorphic to the [[second cohomology group for trivial group action]] <math>H^2(G, A)</math> || {{intermediate notions short|normalized 2-coboundary for trivial group action|normalized 2-cocycle for trivial group action}}
|-
| [[bihomomorphism]] with both input groups identical (equal to <math>G</math>) and the output group abelian (equal to <math>A</math>) || A function <math>f:G \times G \to A</math> such that, if either input for <math>f</math> is fixed, <math>f</math> is a homomorphism from the other input to <math>A</math> || [[bihomomorphism to abelian group implies 2-cocycle]] proves the 2-cocycle part; the normalized part follows directly from the definition of bihomomorphism || || || {{intermediate notions short|normalized 2-cocycle for trivial group action|bihomomorphism}}
|}


<math>0 \to B^2(G, A) \to Z^2(G, A) \to H^2(G, A) \to 0</math>
===Weaker properties===


Imagine downward arrows from the terms of the first short exact sequence to the corresponding terms of the second one. The middle arrow is the identity map; the left arrow (<math>A \to B^2(G, A)</math>) is injective, and the right arrow (<math>Z^2_n(G, A) \to H^2(G, A)</math>) is surjective. Note that the lower short exact sequence does not have a canonical splitting.
{| class="sortable" border="1"
! Property !! Meaning !! Proof of implication !! Proof of strictness !! Capture of difference !! Intermediate notions
|-
| [[Stronger than::2-cocycle for trivial group action]] || || || || The quotient group (with canonical splitting) is isomorphic to the base group (<math>A</math>), with the canonical splitting being constant functions; see [[#Importance]] for more || {{intermediate notions short|2-cocycle for trivial group action|normalized 2-cocycle for trivial group action}}
|}

Latest revision as of 20:24, 7 December 2024

Definition

Equivalent definitions in tabular format

Suppose G is a group and A is an abelian group. We denote by 1 the identity element of G and by 0 the identity element of A; the group operation for G is denoted multiplicatively and the group operation for A is denoted additively. A normalized 2-cocycle for trivial group action is a function f:G×G→A that satisfies the following equivalent conditions:


No. Shorthand Full statement
1 2-cocycle for trivial group action that is zero along the axes f is a 2-cocycle for trivial group action: f(g,hk)+f(h,k)=f(gh,k)+f(g,h)∀g,h,k,∈G and f(g,1)=f(1,g)=0∀g∈G.
2 2-cocycle for trivial group action that is zero at origin f is a 2-cocycle for trivial group action: f(g,hk)+f(h,k)=f(gh,k)+f(g,h)∀g,h,k,∈G and f(1,1)=0.
3 System of coset representatives where identity is represented by identity f corresponds to a system of coset representatives for a central extension E (corresponding to an element of H2(G,A), the second cohomology group for trivial group action) where the coset representative of the identity element is the identity element.

This definition is presented using a tabular format. |View all pages with definitions in tabular format

The set of normalized 2-cocycles for trivial group action forms a subgroup (under pointwise addition) of the group of 2-cocycles for trivial group action. We will denote this subgroup as Zn2(G,A), though this is not standard notation. For the relationship with 2-coboundaries, constant functions, and second cohomology, see #Importance.

Equivalence of definitions

The equivalence of definitions (1) and (2) follows from 2-cocycle for trivial group action is constant on axes, which states that f(g,1)=f(1,g)=f(1,1)∀g∈G.

The equivalence with definition (3) is discussed in more detail at #Relation with central extensions.

Importance

Normalizing a 2-cocycle (finding a normalized 2-cocycle that differs from it via a 2-coboundary)

Translation normalization

The simplest normalization rule is a translation rule: given a 2-cocycle for trivial group action f:G×G→A, we can obtain a normalized 2-cocycle f0:G×G→A defined as follows:

f0(g,h)=f(g,h)−f(1,1)

f0 satisfies the identity for a 2-cocycle because it's just taking the original identity and subtracting f(1,1) twice from both sides.

f0 is also normalized, because f0(g,1)=f(g,1)−f(1,1) and f0(1,g)=f(1,g)−f(1,1), which are both zero based on 2-cocycle for trivial group action is constant on axes.

Moreover, the difference of f and f0 is a constant function G×G→A with constant value f(1,1), which is therefore also a 2-coboundary (of the constant function G→A taking value f(1,1). Thus, f0 belongs to the same cohomology class as f, i.e., they project to the same element of the second cohomology group for trivial group action H2(G,A).

Alternative normalization

Given a 2-cocycle for trivial group action f:G×G→A, we can obtain a normalized 2-cocycle f1:G×G→A defined as follows:

  • f1(g,h)=f(g,h) if g,h,gh are all non-identity elements
  • f1(g,1)=f1(1,g)=0 for all g
  • f1(g,g−1)=f(g,g−1)+f(1,1) for all g (note that it follows from the definition of 2-cocycle that f(g,1)=f(1,1) for all g, so this can also be written as f1(g,g−1)=f(g,g−1)+f(g,1))

f1 is different from f0. As the later discussion will clarify, they differ by a 2-coboundary for trivial group action.

For most real-world purposes, we go with f0. However, f1 could be useful if we want to preserve as many of the values of f as possible.

The group of 2-cocycles is a direct sum of the subgroup of normalized 2-cocycles and the base group (realized as constant functions)

We have:

Z2(G,A)=Zn2(G,A)⊕A

This is a canonical split as an internal direct sum. The terms are as follows:

  • Z2(G,A) is the group of all 2-cocycles for trivial group action.
  • Zn2(G,A) is the subgroup comprising normalized 2-cocycles for trivial group action.
  • A is identified with the subgroup comprising constant functions G×G→A. Specifically, an element a∈A corresponds to the constant function ca sending everything in G×G to a.

To show that this is an internal direct sum, we need to verify the following:

  • Zn2(G,A) is a subgroup of Z2(G,A): This is obvious from the definition.
  • A (identified as constant functions) is a subgroup of Z2(G,A): This is also obvious from the definition of 2-cocycle.
  • Zn2(G,A)∩A=0: For f a normalized 2-cocycle, f(1,1)=0. If f is also constant, this forces it to be zero everywhere.
  • Zn2(G,A)+A=Z2(G,A), i.e., every 2-cocycle can be expressed as the sum of a normalized 2-cocycle and a constant function: We already demonstrated this in the preceding section in the form of translation normalization.

Some implications of this:

  • Z2(G,A)/Zn2(G,A)≅A
  • Z2(G,A)/A≅Zn2(G,A) (where the group A in the denominator is realized as constant functions)
  • 0→Zn2(G,A)→Z2(G,A)→A→0 is a canonically split short exact sequence.
  • 0→A→Z2(G,A)→Zn2(G,A)→0 is a canonically split short exact sequence.

Putting together 2-coboundaries, normalized 2-cocycles, and cocycles

Let's articulate the notation:

  • Z2(G,A) is a group of 2-cocycles for trivial group action.
  • Zn2(G,A) is the subgroup comprising normalized 2-cocycles.
  • B2(G,A) is the subgroup comprising 2-coboundaries.
  • Bn2(G,A) is the subgroup comprising normalized 2-coboundaries; it is the intersection Zn2(G,A)∩B2(G,A).
  • A is identified with constant functions G×G→A.
  • H2(G,A) is the quotient Z2(G,A)/B2(G,A) and is called the second cohomology group for trivial group action.

We have two short exact sequences, the first of which splits canonically and was discussed in the preceding section:

0→A→Z2(G,A)→Zn2(G,A)→0

0→B2(G,A)→Z2(G,A)→H2(G,A)→0

The new fact of interest is that all constant functions are also 2-coboundaries! That's because the constant function G×G→A taking value a∈A can be realized as the coboundary of the constant function G→A taking value a. This means that A maps injectively into B2(G,A). This allows us to make a downward arrow from A to B2(G,A) that is injective. We can also make a downward arrow between the middle groups Z2(G,A) that is the identity map. This in turn allows us to make a downward map from Zn2(G,A) to H2(G,A), and shows that this map must be surjective. In other words, every cohomology class can be represented by a normalized 2-cocycle. The kernel of this map is Bn2(G,A), so we obtain this short exact sequence:

0→Bn2(G,A)→Zn2(G,A)→H2(G,A)→0

We also have a canonical internal direct sum splitting:

B2(G,A)=Bn2(G,A)⊕A

This uses the same logic as the splitting for Z2(G,A) from the preceding section, and the fact that all the constant functions are in fact 2-coboundaries.

We can put all this information together in the commutative diagram below, whose rows are short exact sequences and the last column is also a short exact sequence; the first two rows also have canonical splittings:

(0↓0→A→Zn2(G,A)→Bn2(G,A)→0↓↓↓0→A→Z2(G,A)→Zn2(G,A)→0↓↓↓0→B2(G,A)→Z2(G,A)→H2(G,A)→0↓0)

Relation with treatment as central extensions

This subsection looks at 2-cocycles in terms of their interpretation in terms of central extensions. It leads to a restatement / rediscovery / different perspective on some of the results of the preceding subsections.

Let E be a group with a central subgroup isomorphic to (and explicitly identified with) A, and a quotient isomorphic to (and explicitly identified with) G, such that the induced action of the quotient on the subgroup (in the sense of action by conjugation, see quotient group acts on abelian normal subgroup). Let S be a system of coset representatives for G in E with s:G→S being the representation map. Then, define f:G×G→A such that

s(gh)=f(g,h)s(g)s(h)

In other words, f measures the extent to which the collection of coset representatives fails to be closed under multiplication.

Such an f is a 2-cocycle for trivial group action of G on A. The different possible 2-cocycles for a given central extension form a single coset of the subgroup of 2-coboundaries, and hence a single element of the second cohomology group for trivial group action H2(G,A).

The key insight is as follows:

The 2-cocycle is a normalized 2-cocycle if and only if s(1) (which needs to be an element of A inside E) is the identity element of E (and hence also of A as a subgroup of E).

Since it's always possible to choose s so that s(1) is the identity element of A, every central extension can be represented by a normalized 2-cocycle. In other words, every element of the second cohomology group for trivial group action H2(G,A) can be represented by a normalized 2-cocycle. This is just a restatement of our earlier observation that Zn2(G,A) maps surjectively to H2(G,A).

We can explain the two normalizations discussed above as follows:

  • The translation normalization corresponds to taking the original system of coset representatives and multiplying all of them by s(1)−1 (left and right doesn't matter because s(1) is central).
  • The alternative normalization corresponds to keeping all coset representatives intact except s(1), which is replaced by the identity element.

Relation with other properties

Stronger properties

Property Meaning Proof of implication Proof of strictness Capture of difference Intermediate notions
cyclicity-preserving 2-cocycle for trivial group action A 2-cocycle f such that f(g,h)=0 whenever ⟨g,h⟩ is cyclic IIP 2-cocycle for trivial group action|FULL LIST, MORE INFO
IIP 2-cocycle for trivial group action A 2-cocycle f such that f(g,g−1)=0 for all g |FULL LIST, MORE INFO
normalized 2-coboundary for trivial group action A 2-coboundary f such that f(g,1)=f(1,g)=0 for all g The quotient group is canonically isomorphic to the second cohomology group for trivial group action H2(G,A) |FULL LIST, MORE INFO
bihomomorphism with both input groups identical (equal to G) and the output group abelian (equal to A) A function f:G×G→A such that, if either input for f is fixed, f is a homomorphism from the other input to A bihomomorphism to abelian group implies 2-cocycle proves the 2-cocycle part; the normalized part follows directly from the definition of bihomomorphism |FULL LIST, MORE INFO

Weaker properties

Property Meaning Proof of implication Proof of strictness Capture of difference Intermediate notions
2-cocycle for trivial group action The quotient group (with canonical splitting) is isomorphic to the base group (A), with the canonical splitting being constant functions; see #Importance for more |FULL LIST, MORE INFO