How to Define Homomorphisms on Quotient Groups
By Mu Prime Math
Summary
## Key takeaways - **Defining maps on quotient groups needs care**: When defining a homomorphism f from a quotient group G mod K to a group H, using coset representatives from G can be problematic. The function f must be independent of the chosen representative for it to be well-defined. [00:34], [02:49] - **Integers mod 5 example shows well-definedness failure**: An attempt to define a homomorphism from Z mod 5Z to Z using the identity map on integers fails because different representatives of the same coset (e.g., 2 and 7 in 2+5Z) map to different values (2 and 7, respectively). [03:39], [05:20] - **Well-definedness condition: map must be identity on normal subgroup**: For a homomorphism phi from G to H to induce a well-defined homomorphism on G mod K, phi must map every element of the normal subgroup K to the identity element in H. [06:43], [10:09] - **Alternating group example: sign homomorphism works**: The sign homomorphism from the symmetric group SN to {+1, -1} is well-defined on the quotient SN mod An because all elements in the normal subgroup An map to +1 (the identity). [11:20], [12:40] - **Quotients are key in abstract algebra constructions**: Many important mathematical objects, like tensor products, are constructed as quotient spaces. Defining maps on these quotients often involves verifying properties on the original space and the subspace being factored out. [13:31], [14:13]
Topics Covered
- Why defining quotient homomorphisms isn't as simple as it seems.
- Every subgroup element must map to the identity.
- The sign permutation map works for alternating groups.
- How quotient homomorphisms enable complex mathematical constructions.
Full Transcript
so we are going to talk about defining
homomorphisms on quotients this video is
going to use the language of groups but
the same idea applies to other algebraic
structures such as rings
so let's say we have a group G and a
normal subgroup K of G we can form the
quotient Group G mod K and the elements
of that quotient are cosets which look
like G times K where little G is an
element of the original group
now let's say we want to define a
homomorphism f that goes from the
quotient Group G mod K to some other
group h
one way we can Define such a
homomorphism is to say well F the
elements of the quotient group The
Domain of f
those are these cosets G times k
and when we take a coset this element
here this little G is called a
representative of the coset G times K
it's a specific element of the coset
that we choose to represent the entire
set
so one way to define F of G times K is
to just look at the coset representative
and Define f as a function of the
representative
so to do this we're essentially defining
another homomorphism fee that goes from
the original group G to that same
codomain H and then we're using this fee
to Define f in terms of the coset
representatives
now if we Define f in this way if Phi is
a group homomorphism then F defined like
this is automatically also a group
homomorphism because if we look at F of
g 1 K times G2K where G1 and G2 are
elements of that group
in here this g1k times G2K is the same
as G1 G2K that's how multiplication in
the quotient group is defined
and then
by definition F of G1 G2K that's going
to be Phi of G1 G2
and then on the other hand F of g1k
times F of G2K
well we can compute each of these
individually using this definition so F
of g 1 K is Phi of G1 and similarly F of
G2K is Phi of G2 now because Phi is a
group homomorphism Phi of G1 G2 equals
Phi of G1 times Phi of G2 which means
that these two first expressions are
equal and therefore f is a group
homomorphism
so it seems like we're done if we want
to define a homomorphism on the quotient
group gmod K all we have to do is find a
homomorphism on the original group G and
then use that to define a function on
the cosets by just picking a coset
representative and plugging that into
Phi
there's one problem here though which is
we don't know whether the function f is
well defined
to see an example of where this problem
comes from let's take a look at the
group of integers under addition
we can define a group homomorphism from
the integers to Itself by just taking
Phi of n equals n this is the identity
map on the integers of course it's a
group homomorphism
now if we take a look at the subgroup 5z
this is a normal subgroup of the entire
group of integers Z under addition so
this subgroup here is just the set of
all integers that are multiples of five
and we can construct the quotient group
Z mod 5z so this is the integers mod 5.
and using this construction we could
Define a function f that goes from Z mod
5z to the same codomain Z
and the way we want to Define that is by
saying F of n plus 5z this is our coset
and the definition here is that we just
take the coset representative and apply
Phi to that so Phi of n
but Phi of n is just equal to n
so take a second and see if you can
figure out the problem with this
definition of a function f of n plus 5z
equals n
the problem here is that
n plus 5z is a coset that contains many
different elements so as an example
let's take a look at the coset 2 plus 5z
so this coset contains 2 plus every
integer multiple of five so the elements
are going to look like 2 7 12 17 and
then also in the reverse Direction we'd
have minus three minus 8 minus 13 and so
on
but the key Point here is that 2 plus 5z
is the same coset as 7 plus 5z because
if we take 2 and add all the multiples
of 5 the set that we get is the same set
that we would get if we start with 7 and
then add all the multiples of five
but if these are the exact same set then
we need to have F of 2 plus 5z being
equal to F of seven plus five Z
and that's not what we get with this
definition because F of 2 plus 5z is
just 2 and F of 7 plus 5z is seven but
that doesn't make any sense because
these 2 plus 5z and 7 plus 5z they're
the same set so we can't have a function
that Maps the same set to two different
outputs
so let's take this problem back to the
more General language of an arbitrary
group
the problem here is that if we have some
coset GK of a normal subgroup k then we
can also have a different coset
representative let's say the first
representative is G1 and the second
representative is G2 so we're almost
always going to have cases in a quotient
group where there are two elements G1
and G2 of that original group where g1k
equals G2K and if that's the case then
we need F of g1k to equal F of G2K
if we want to define a homomorphism on
the quotient group in terms of
homomorphism on the original group G
then we're having F of G1 K being equal
to Phi of G1 and then F of G2K is Phi of
G2
so in order for the function f to be
well defined we need to have that if g1k
equals G2K then Phi of G1 equals Phi of
G2 so let's think a little bit more
about what this equation here means if
g1k equals G2K one thing we can do here
is multiply both sides of the equation
by G2 inverse on the left
so on the left side we'll get G2 inverse
G1 K and then on the right side G2
inverse G2K is just k
now both sides of these equation these
are sets of elements in the original
group G so of course this K on the right
side is just that normal subgroup but on
the left side G2 inverse g1k that's the
set of all elements of the form G2
inverse times G1 times some element of
the normal subgroup k
now in particular K contains the
identity Element e of the original group
so G2 inverse G1 times the identity
element
this is in G2 inverse G1 K this is part
of that coset
and we know that G2 inverse g1k equals
that normal subgroup k
which means that G2 inverse G1 times the
identity is an element
of K of course the identity element
doesn't do anything when we multiply so
this is telling us that G2 inverse G1 is
an element of K so this first equation
g1k equals G2K
that means that G2 inverse G1 is in that
normal subgroup k
now let's take a look at the second
equation over here Phi of G1 equals Phi
of G2
now in this case Phi of G1 and Phi of G2
these are both elements of the second
group h
so what I'm going to do is multiply by
the inverse of this element Phi of G2 on
both sides
so on the left side we're going to get
Phi of G2 inverse times Phi of G1 and
then on the right side Phi of G2 inverse
times Phi of G2 that's just the identity
element
but fee is a group homomorphism so first
of all Phi of G2 and then taking the
inverse that's the same as Phi of G2
inverse
and then second of all Phi of G2 inverse
times Phi of G1 is the same as Phi of G2
inverse G1 so those are just basic
properties of the group homomorphism
so Phi of G2 inverse G1 equals the
identity that's what this second
equation is telling us
so the new form we have of the
implication that we're looking for here
is that if G2 inverse G1 is in k then
Phi of G2 inverse G1 equals the identity
but notice in both cases here that this
element here G2 inverse G1 that's the
same as what we're plugging into Phi
so what we're really saying is that if
we have some element of the group G
that's in the normal subgroup k then Phi
has to be the identity on that element
of the group that's the thing that we
want in order for the homomorphism to be
well defined
so in other words fee applied to any
element of the normal subgroup K has to
be the identity
so this is the condition that we need to
check whether a homomorphism on the
original group G gives us a well-defined
homomorphism on the quotient Group G mod
k
we start with any homomorphism fee from
G to H and all we need to check is that
Phi gives us the identity on every
element of that subgroup k
as an example of this condition we can
take a look at the homomorphism fee from
the integers to itself given by Phi of n
equals n
in this case we're considering the
normal subgroup 5z that's basically our
K in this case and one element of 5z
that's all the multiples of 5 is just
the integer five but Phi of five equals
five and that's not the identity element
of the integers it's not zero so this
homomorphism fee does not satisfy this
condition which is why it doesn't give
us a well-defined homomorphism on the
quotient group Z mod 5z
next I'll give an example of this
condition working if you're familiar
with the symmetric group and the sign of
a permutation
so there's a group homomorphism Sigma
that goes from the symmetric group SN to
the set plus or minus one this set is a
group under multiplication and the
homomorphism is defined by Sigma of some
permutation g equals the sine of G
so the sign of a permutation tells us
the number of two cycles that are needed
to create the permutation G so if the
number of two cycles that we need to
create the permutation is even then the
sine of G is plus one and we say that g
is an even permutation whereas if the
number of two cycles that we need is odd
then we say that g is an odd permutation
and its sine is negative one so for
example the permutation one two three
this is a three cycle but we can write
it as 1 2 times 2 3. so both one two and
two three are two cycles so this is
writing the permutation as an even
number of two cycles which means that
this is an even permutation
now
the set of even permutations in the
symmetric group actually forms a normal
subgroup it's called the alternating
group a n
so it's a subgroup of index two
and the point is that the sine of every
element in the alternating group is plus
one because by definition this is the
set of all even permutations so it's the
set of all things with sine one and plus
one is the identity element of this
group here
so every element of this normal subgroup
gets sent to the identity by this map
Sigma which means that we can use Sigma
to define a map F on the quotient SN mod
a n to the group plus or minus one if we
Define f on some permutation G times a n
this is a coset of the alternating group
if we Define this as just Sigma of that
permutation G then this gives us a
well-defined homomorphism from SN mod a
n to the group plus or minus one
now the reason that this idea of
defining homomorphisms on a quotient is
so important is that there are many
important objects in math that are best
defined as quotient groups quotient
spaces quotient modules or something
like that
for example if you're familiar with the
tensor product of vector spaces that is
something that we often construct as a
quotient space of a very big Vector
space and then we take the quotient by a
Subspace generated by a bunch of
different elements
so if we Define the tensor product as
this quotient let's say a mod b
it can be very difficult to check
whether a map from a mod b to some other
Vector space is well defined because we
don't know whether it's independent of
the coset representative that we choose
here
so a lot of the time instead of defining
a map directly on the tensor product
what we do is start by defining a map on
the original group a and then we verify
that that map is zero on every element
of this Subspace B that we use to take
the quotient because if the map on a is
zero on every element of B then this
automatically tells us that we get a
well-defined map from a mod b to
whatever other Vector space we're
interested in
so this is a key way that we can Define
homomorphisms on quotients
[Music]
Loading video analysis...