Let $G$ be $l$-dimensional compact Lie group and consider any smooth $F : \mathbb{R}^n \to G$.

Then, the first-order derivative of $F$ at each $x \in \mathbb{R}^n$ can be regarded as a linear mapping $D_x F : \mathbb{R}^n \to \mathbb{R}^l$. Let us fix a matrix norm $\lVert \cdot \rVert$ and assume that \begin{equation} x \mapsto \lVert D_x F \rVert \text{ is polynomially bounded} \end{equation} We can extend the above notion of polynomial boundedness to higher order derivatives as well.

Now, let us define a set $\mathfrak{G}$ as the set of $F : \mathbb{R}^n \to G$ such that $F$ is smooth and all its derivatives are polynomially bounded in the above sense.

Now, my question is that

Is this $\mathfrak{G}$ a group under pointwise multiplication and inverse?

That is,

  1. For any $F_1, F_2 \in \mathfrak{G}$ and $x \in \mathbb{R}^n$, let $(F_1 F_2)(x) := F_1(x) F_2(x)$. Then, do we have $F_1 F_2 \in \mathfrak{G}$?

  2. For any $F \in \mathfrak{G}$ and $x \in \mathbb{R}^n$, let $F^{-1}(x) := [F(x)]^{-1}$. Then, do we have $F^{-1} \in \mathfrak{G}$.

I think the item 1 must be true, but not sure about item 2. For example, can a derivative of $F$ decay exponentially at infinity, making a derivative of $F^{-1}$ "grow exponentially" at infinity?

Could anyone please clarify for me? If I am wrong about $\mathfrak{G}$ being a group, what additional conditions should I impose to make it into a group?

Add) I think this is indeed the case for $\operatorname{SU}(N)$ with $N \geq 2$ or any direct product of them. Still, I haven't figured out for general $G$.

Add 2) This ME post shows that my guess is also true for $\operatorname{U}(1)$.


1 Answer 1


Note that the maps $$\iota: G\to G,\ x\mapsto x^{-1}\qquad \mu:G\times G\to G,\ (x,y)\mapsto x\cdot y$$ have bounded derivatives, since you are assuming $G$ to be compact.

Then for $F_1, F_2\in\mathfrak G$ you have that the derivatives of $$F_1^{-1}=\iota\circ F_1,\qquad F_1\cdot F_2 = \mu\circ (F_1,F_2)$$ will have polynomial growth, being the composition of a map whose derivative has polynomial growth and a map with bounded derivative.

  • $\begingroup$ I do not clearly see how we can use compactness of $G$ to show that $\iota: G\to G$ has bounded derivatives. Could you provide more details? $\endgroup$
    – Isaac
    Commented Jul 16 at 22:44
  • $\begingroup$ $d\iota: G \to \Bbb R^\ell$ is a continuous function with compact domain. $\endgroup$
    – mme
    Commented Jul 17 at 0:18
  • $\begingroup$ @mme I am confused about the domain and range of your differential. Are you thinking pointwise on $G$ or the whole tangent bundle? $\endgroup$
    – Isaac
    Commented Jul 17 at 9:48
  • 1
    $\begingroup$ @Isaac the norms $\|d\iota\|$ and $\|d\mu\|$ are continuous $\Bbb R$-valued functions on $G$ and $G\times G$. Compactness of $G$ guarantees that their image is bounded. $\endgroup$
    – o r
    Commented Jul 17 at 15:30
  • 1
    $\begingroup$ Yes, I mean that function. The argument can be applied to show the boundedness of the norms of the higher order derivatives of $\iota$ and $\mu$ as well (higher derivatives are usually described via "jets" in differential geometry). $\endgroup$
    – o r
    Commented Jul 17 at 17:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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