Let $X$ be a smooth manifold. What is the appropriate topology on $C^\infty(X)$ such that a linear functional $\lambda$ on $C^\infty(X)$ is continuous iff it can be represented as a limit of the form $$\lambda f = \lim_{\ell\to\infty} \sum_{k=1}^{n_\ell}\lambda_{k\ell}f(x_{k\ell})$$ with finite $n_\ell$, suitable numbers $\lambda_{k\ell}$, and suitable points $x_{k\ell}\in X$?

  • $\begingroup$ As far as I understand, the usual topology of the locally convex projective limit $C^\infty(X)=\underset{U}{\lim} C^\infty(U)$, $U\cong {\mathbb R}^n$ does not satify you? $\endgroup$ – Sergei Akbarov Jul 3 '18 at 13:15
  • $\begingroup$ @SergeiAkbarov: Maybe it would satisfy me if the topology were expressed in a less abstract way and if I'd see the argument why my statement is then true. $\endgroup$ – Arnold Neumaier Jul 3 '18 at 13:40

Arnold, if $l$ is not obliged to belong to $\mathbb N$, i.e. the index set for $l$ can be arbitrary (for $l\in\mathbb N$ this seems to be also true, but this requires a verification), then the usual (the weakest locally convex) topology of $C^\infty(M)$ generated by the projections to the spaces $C^\infty(U)$, where $U$ is diffeomorphic to ${\mathbb R}^n$ (or to an open subset in ${\mathbb R}^n$, the topology on such $C^\infty(U)$ is described in the Kirillov-Gvishiani book), satisfies your requirement, since the linear combinations of delta-functionals are dense in $(C^\infty(M))'$ (this in its turn follows from the fact that delta-functionals separate elements of $C^\infty(M)$).

In this topology a net of functions $f_\nu$ tends to a function $f$ in $C^\infty(M)$ if and only if for each subset $U\subseteq M$, diffeomorphic to $\mathbb R^n$, the restrictions $f_\nu\big|_U$ tend to $f\big|_U$ uniformly on each compact set $K\subseteq U$ by each partial derivative $\partial^m$ generated by the local chart for $U$: $$ \forall m\in{\mathbb N}^n\qquad \sup_{x\in K}\bigg|\partial^m\Big(f_\nu\big|_U-f\big|_U\Big)(x)\bigg|\underset{\nu\to\infty}{\longrightarrow}0 $$ (I assume everywhere that $\mathbb N$ contains zero).

There is another "abstract way" to describe this topology that I find "more visual", I told about it here:

  1. For each function $f\in {\mathcal C}^\infty(M)$ let us define its support as the closure of the set of the points where $f$ does not vanish: $$ \text{supp}f=\overline{\{x\in M:\ f(x)\ne 0\}}. $$ An equivalent definition: $\text{supp}f$ is the set of the points in $M$ where $f$ has non-zero germs: $$ \text{supp}f=\{x\in M:\ f\not\equiv 0\ (\text{mod}\ x)\}. $$

  2. Let us define differential operators (see e.g. S.Helgason's book) on $M$ as linear mappings $D:{\mathcal C}^\infty(M)\to {\mathcal C}^\infty(M)$ which do not extend the support of functions: $$ \text{supp}Df\subseteq \text{supp}f,\quad f\in{\mathcal C}^\infty(M). $$ Equivalently, $D$ is local, i.e. the value of $Df$ in a point $x\in M$ depends only on the germ of $f$ in $x$: $$ \forall f,g\in{\mathcal C}^\infty(M)\quad \forall x\in M\qquad f\equiv g\ (\text{mod}\ x)\quad\Longrightarrow\quad Df(x)=Dg(x). $$

  3. Then we say that a sequence of functions $f_n$ converges to a function $f$ in ${\mathcal C}^\infty(M)$ $$ f_n\overset{{\mathcal C}^\infty(M)}{\underset{n\to\infty}{\longrightarrow}}f $$ if and only if for each differential operator $D:{\mathcal C}^\infty(M)\to {\mathcal C}^\infty(M)$ the sequence of functions $Df_n$ converges to $Df$ in the space ${\mathcal C}(M)$ of continuous functions with the usual topology of uniform convergence on compact sets in $M$: $$ Df_n\overset{{\mathcal C}(M)}{\underset{n\to\infty}{\longrightarrow}}Df $$

The definition of the differential operator in this construction belongs (as far as I know) to Jaak Peetre. We can replace it by the one that Jet Nestruev uses for abstract rings, and the topology on $C^\infty(M)$ will be the same. Similarly, we can replace $D$ by arbitrary compositions of vector fields on $M$ $$ D=X_1\circ...\circ X_p, \quad p\in {\mathbb N}, $$ where vector fields are defined as derivations of the ring $C^\infty(M)$, i.e. linear operators $X:C^\infty(M)\to C^\infty(M)$ with the Leibnitz property $$ X(f\cdot g)=X(f)\cdot g+f\cdot X(g). $$

| cite | improve this answer | |
  • $\begingroup$ I like the final definition under 3. How to see from that definition that the linear combinations of delta-functionals are dense in (C∞(X))′ without first proving equivalence to the first definition? $\endgroup$ – Arnold Neumaier Jul 3 '18 at 14:35
  • $\begingroup$ Arnold, I doubt that this can be seen without proving the equivalence. In my head the explanation is the following. First, we prove that $C^\infty(M)$ is a Fréchet space for $\sigma$-compact $M$. Second, we prove that for arbitrary $M$ the space $C^\infty(M)$ is isomorphic to the direct product of a family of spaces $C^\infty(M_i)$ with $\sigma$-compact $M_i$. Together this means that $C^\infty(M)$ is always stereotype. After that we conclude that the bipolar theorem (see Schaefer IV, 1.5) holds for the pair $\Big(C^\infty(M),C^\infty(M)^\star\Big)$. $\endgroup$ – Sergei Akbarov Jul 3 '18 at 15:20
  • $\begingroup$ As a corollary, if a subspace $P\subseteq C^\infty(M)^\star$ has zero polar in $C^\infty(M)$, then its bipolar $P^{\circ\circ}$, which is the $C^\infty(M)$-weak closure of $P$, and thus the closure of $P$ in $C^\infty(M)^\star$, coincides with $C^\infty(M)^\star$. In other words, $P$ must be dense in $C^\infty(M)^\star$. Finally, we take $P$ as the linear span of delta-functions, and we obtain what we need. $\endgroup$ – Sergei Akbarov Jul 3 '18 at 15:32
  • $\begingroup$ The only reference that comes to my mind is Kirillov-Gvishiani springer.com/gb/book/9780387906386 But they consider only the case when $M=U$ is an open subset in ${\mathbb R}^n$. However, from their Theorem 26 in Chapter III (where they prove that $C^\infty(U)$ which they denote by ${\mathcal E}(U)$ is a Fréchet space) it follows what you need, you should just follow the scheme that I described. $\endgroup$ – Sergei Akbarov Jul 3 '18 at 17:54
  • $\begingroup$ I understand. I think there must be a textbook where all the details are explained accurately. We should ask people to give a reference. $\endgroup$ – Sergei Akbarov Jul 3 '18 at 22:30

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.