2024-10-15

What is a group?

A group is , where

  • is a set.
  • .
  • .
  • .

There are a bunch of axioms.

Associative law

G£G£GG£GG£GGid£m23m12£idmmg1(g2g3)=(g1g2)g3.

Identity element

G£¤G£GGid£epr1mge=g.
¤£GG£GGid£epr2meg=g

Inverses

Let be the diagonal map.

G£GG£GG¤Gid£im¢egg¡1=e.
G£GG£GG¤Gi£idm¢eg¡1g=e.

Definition (Group object)

A group object in a category is …

Let’s first specify what should have.

The category should have

  • Products. That is, for all , there exists satisfying the universal property
WX£YYX9!pr2pr1
  • A terminal object such that is a one point set.
W¤9!

Definition (Group object, v1)

Assume has products and a terminal object. A group object in a category is such that the above are satisfied.

Examples

  • In , we get .
  • In , we get .
  • In , we get .
  • In , we get .
  • We can also consider .
  • In , a field, we get .

Alternative description using Yoneda embeddings

Let and define the “functor of points”

If is a group object, then is a group (and vice versa). That is, the functor factors through . For example,

Definition 2/Theorem

A group object is such that factors through .

Example

In , we have

Some facts about :

  • The product in is the tensor product of rings.
  • The terminal object is .

We have

This is a group under matrix multiplication. (For , we get a group scheme, for a field, we get an algebraic group.)

Tangent spaces

In Lie groups, group schemes, algebraic groups, we can talk about tangent spaces. (Algebraic groups are smooth.)

In cases where there is a tangent space, the Lie algebra of is

the tangent space at the identity. Often written with .

Example:

The identity is .

In vector calculus, we learn that the tangent space is orthogonal to , where the space is given by . Here, our function is

So the tangent space is given by

So tangent spaces are “easy”.

It looks like it uses calculus, but it doesn’t actually use calculus when the functions we use are polynomials.

Curves in

Let . A curve in is given as

Then is a tangent vector. If we put in the condition that , then we have

  • Constant term: .
  • Linear term: . This gives the same as above.

Notice that this was a completely algebraic computation: we didn’t care about terms or higher. This is really a computation in the ring .

In an algebraic setting

Let be an algebraic group (over a field). Then

where

Alternatively, we are looking at the composition

Example:

As a functor,

The Lie algebra

To see this, is fairly easy to compute since is linear in each variable. The entries in are determinants of matrices of the submatrices. At the identity, the entries of are all (for on-diagonal entries) or (for off-diagonal entries). The trace is the sum of the diagonal entries.

Remark