Following the definition of $\text{GL}_n:\mathbf{Alg}_k\rightarrow\mathbf{Set}:R\mapsto \text{GL}_n(R)$ found here: functor from $\mathbf{Alg}$ to $\mathbf{Set}$

I would like to show that it is a representable functor.

EDIT: wrong!

For ease of notation, $F:=\text{GL}_n$ and $G:= \text{Hom}(k,-)$

Where $\text{Hom}(k,-):\mathbf{Alg}_k\rightarrow\mathbf{Set}:R\mapsto \text{Hom}(k,R):f\mapsto (\sigma\mapsto f\circ \sigma)$

I'll define the natural transformation as:

$$\alpha_R:\text{GL}_n(R)\rightarrow \text{Hom}(k,R):(a_{i,j})\mapsto (k\mapsto \sum a_{i,j}k)$$

this defines a natural transformation. But how natural is it? I.e. are there any more intuitive ways of doing this?

EDIT: answer (from answer below and adapted to my notations)

For ease of notation, $F:=\text{GL}_n$ and $G:= \text{Hom}(A_n,-)$

where $$A_n=k[X_{11},\ldots,X_{nn},Y]/\left<1-Y\det(X_{ij})\right>.$$

Where $\text{Hom}(A_n,-):\mathbf{Alg}_k\rightarrow\mathbf{Set}:R\mapsto \text{Hom}(A_n,R):f\mapsto (\sigma\mapsto f\circ \sigma)$

I'll define the natural transformation as:

$$\alpha_R:\text{GL}_n(R)\rightarrow \text{Hom}(A_n,R)\\(a_{i,j})\mapsto (P(X_{11},\ldots,X_{nn},Y)+ (1-Y\det(X_{ij}))\mapsto P(a_{1,1},\ldots,a_{n,n},\text{det}(a_{i,j})))$$

  • $\begingroup$ It will certaibly not be represented by $k$. Do you think $GL_n(k) \simeq Hom(k,k)$ ? $\endgroup$ – Maxime Ramzi Dec 30 '17 at 11:52
  • $\begingroup$ Oups, I forgot to check that it is a bijection. What I gave is a natural transformation but not necessarily a natural isomorphism. $\endgroup$ – tomak Dec 30 '17 at 12:00
  • $\begingroup$ Only other idea is to represent it by $k^n$ but not sure that works $\endgroup$ – tomak Dec 30 '17 at 12:01
  • $\begingroup$ There's the same problem with $Hom(k^n, k) \simeq GL_n(k)$ $\endgroup$ – Maxime Ramzi Dec 30 '17 at 12:40

Yes, $\text{GL}_n$ is an affine group scheme. It is represented by the $k$-algebra $$A_n=k[X_{11},X_{12},\ldots,X_{21},\ldots,X_{nn},Y]/\left<1-Y\det(X_{ij})\right>.$$ An algebra homomorphism $\phi:A_n\to R$ is given by choosing the images of $X_{ij}$ and $Y$, which must go to the inverse of $\det(\phi(X_{ij}))$ which ensures that $(\phi(X_{ij}))\in\text{GL}_n(R)$.

Waterhouse's book is a good introduction to affine group schemes.

| cite | improve this answer | |
  • $\begingroup$ I modified my answer. Can you check that it makes sense? $\endgroup$ – tomak Dec 30 '17 at 14:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.