This post is not about finding an answer to a certain problem - because the answer already exists - but rather about finding the simplest possible answer.

The problem is: how to define the bundle $C(M)$ whose sections are all the symmetric affine connections on the manifold $M$ in terms of natural bundles over $M$ in the simplest possible way?

If we drop the hypotesis of symmetricity, then the answer probably is $$ J^1(F(M))/GL_n(\mathbb{R})\quad\quad (*) $$ where $F(M)$ is the frame bundle. I believe that formula $(*)$ actually becomes simpler with the hypotesis of symmetricity.

My idea goes like this. Consider the following bundle $$ \check{J}^k(M,\mathbb{R}_0^n):=\{[\boldsymbol{x}]_{m,0}^k\mid\boldsymbol{x}\textrm{ is a local diffeomorphism sending }m\in M\textrm{ to }0\in \mathbb{R}^n\} $$ of $k$-jets of local diffeomorphisms from $M$ to $\mathbb{R}^n$ together with its natural projection on $M$ (like in P. Michor's book about manifolds of differentiable mappings). In particular, $$ \check{J}^1(M,\mathbb{R}_0^n)=F(M) $$ is a $GL_n(\mathbb{R})$-principal bundle and each $\check{J}^k(M,\mathbb{R}_0^n)$ carries a natural $GL_n(\mathbb{R})$-action. I'm sure that the sub-bundle of $(*)$ made of symmetric affine connection is precisely $$ C(M)=\check{J}^2(M,\mathbb{R}_0^n)/GL_n(\mathbb{R})\quad\quad (**) $$

QUESTION: is the identification $(**)$ correct? If yes, is there a direct proof of it, i.e., not passing through $(*)$? Can $(**)$ be found, formalised exactly as above, in the literature?

Needless to say, either if my guess $(**)$ is wrong or if there are better answers out there, I'd like to know!


There is an 'identification', i.e., a way to interpret a torsion-free affine connection on $M$ as a section of $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})$ in such a way that every (smooth) section of $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})$ over $M$ corresponds to a unique torsion-free (smooth) affine connection. You haven't explicitly described such an identification, though, so it's hard to say whether $(**)$ is 'correct'.

Here is one way to do it: Given a torsion-free connection $\nabla$ on an $n$-manifold $M$, and a point $m\in M$ let $\exp_m^\nabla:T_mM\to M$ be the (locally defined) exponential map, which is a diffeomorphism from an open neighborhood of $0_m\in T_mM$ to an open $m$-neighborhood $U_m\subset M$. Choose a linear isomorphism $u: T_mM\to\mathbb{R}^n$ and let $x = u^{-1}\circ (\exp_m^\nabla)^{-1}:(U_m,m)\to(\mathbb{R}^n,0)$. Then the $\mathrm{GL}_n(\mathbb{R})$ equivalence class $[x]^2_{m,0}{\cdot} \mathrm{GL}_n(\mathbb{R})$ is a canonically-determined element $\gamma(\nabla)_m$ in the $m$-fiber of the bundle $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})\to M$. Thus, $\nabla$ canonically determines a section $\gamma(\nabla)$ of $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})\to M$.

It's easy to check (in local coordinates) that this mapping from connections to sections of the given bundle has all of the desired properties. In particular, every smooth section $\sigma$ of $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})\to M$ is of the form $\sigma = \gamma(\nabla)$ for some unique smooth torsion-free affine connection on $M$.

Note that the natural affine structure on $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})\to M$ is described as follows: Let $[x]^2_{m,0}{\cdot}\mathrm{GL}_n(\mathbb{R})$ and $[y]^2_{m,0}{\cdot}\mathrm{GL}_n(\mathbb{R})$ be two elements in the $m$-fiber and suppose that two local coordinate representatives have been chosen so that $[x]^1_{m,0}=[y]^1_{m,0}$. Then define $$ t\,\bigl([x]^2_{m,0}{\cdot}\mathrm{GL}_n(\mathbb{R})\bigr)+(1{-}t)\,\bigl([y]^2_{m,0}{\cdot}\mathrm{GL}_n(\mathbb{R})\bigr) = \bigl([t\,x + (1{-}t)\,y]^2_{m,0}{\cdot}\mathrm{GL}_n(\mathbb{R})\bigr). $$ One can check (again, in local coordinates), that this is a smooth affine action on $\check J^2(M,\mathbb{R}^n_0)/\mathrm{GL}_n(\mathbb{R})\to M$ and that $$ \gamma(t\,\nabla_1 + (1{-}t)\,\nabla_2) = t\,\gamma(\nabla_1) + (1{-}t)\,\gamma(\nabla_2). $$

As far as references go, I don't know for sure, but I would not be at all surprised to find that this precise construction is described somewhere in Charles Ehresmann's original papers on natural jet bundles.

Added remark: One way to make this match a little better with the description of the general affine connection is to use, instead, the coframe bundle $F^*(M)\to M$, where a coframe $u\in F^*(M)$ is a linear isomorphism $u:T_mM\to\mathbb{R}^n$. A local section of $F^*(M)$ is just a coframing, i.e., an $\mathbb{R}^n$-valued $1$-form $\eta$ on $U\subset M$ such that $\eta_m:T_mU\to\mathbb{R}^n$ is an isomorphism for all $m\in U$. Then, in a natural way, the space of affine connections is identified with the sections of $J^1(F^*(M))/\mathrm{GL}_n(\mathbb{R})$, while the space of torsion-free affine connections is identified with the space of sections of the $\mathrm{GL}_n(\mathbb{R})$-quotient of the submanifold $J^1_0(F^*(M))\subset J^1(F^*(M))$ consisting of $1$-jets of closed coframings, i.e., the local coframings $\eta$ that satisfy $\mathrm{d}\eta=0$.

  • $\begingroup$ Of course you're right. My reasoning (which I didn't post) was the following: define $\gamma(\Delta)_m$ directly by letting $[x]^2_{m,0}\cdot\mathrm{GL}_n(\mathbb{R})$, where $x:U_m\to\mathbb{R}^n$ is an inertial reference frame at $m$, that is a coordinate system where the Christoffel symbols of $\Gamma$ vanishes at $m$. I think this "my" definition is equivalent to yours. And now I'm sure things work out the way they should. Still, it's frustrating not being able to find such a trifle in the literature. "Folklore" cannot be cited! I'll leave the post open, maybe somebody has a reference $\endgroup$ – Giovanni Moreno Apr 7 '18 at 15:36
  • $\begingroup$ Have you tried looking at Ehresmann's papers? I certainly expect that either it is done somewhere there or a reference is given. $\endgroup$ – Robert Bryant Apr 7 '18 at 17:16
  • 1
    $\begingroup$ In case you want to look through Ehresmann's collected works to try to find a reference there, they are available online for free at ehres.pagesperso-orange.fr/C.E.WORKS_fichiers/C.E_Works.htm $\endgroup$ – Robert Bryant Apr 7 '18 at 17:43
  • $\begingroup$ Excellent. Now I only need a team of french-speaking post-docs working around the clock trying to find a needle in a haystack. $\endgroup$ – Giovanni Moreno Apr 7 '18 at 19:34
  • 3
    $\begingroup$ If it were me having to use this fact in a paper or a lecture, I would simply put in a remark describing the identification (as I did) and say (as I did) that the reader can readily check, via local coordinates, that it has the desired properties. I might add a footnote along the lines of "I am not sure to whom this construction should be attributed, and I would be interested to know a reference to its first explicit use." While the construction is of historical interest, I think it's not a significant enough result to warrant a lot of concern about attribution. $\endgroup$ – Robert Bryant Apr 7 '18 at 21:23

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.