The universality of the spaces of ultradistributions \({\mathcal C}_ S({\mathbb{T}})^\wedge\), \({\mathcal C}_{(S)}({\mathbb{T}})^\wedge\) (\(0<s\leq \infty\)), \({\mathcal C}_ 0({\mathbb{T}})^\wedge\) and \(\text{Exp}({\mathbb{C}}^\times)^\wedge\). (Q1106427)

If Y and E are complete l.c. spaces, Y with approximation property, then \[ Y{\check \otimes}_{\epsilon}E\simeq Y\epsilon E=L_ e(Y_ c',E). \] For many l.c. spaces \({\mathcal C}(T)\) of functions on the unit circle T (like C(T) or \(C^{\infty}(T))\) this has the following interpretation: For each function u:T\(\to E\) satisfying y'\(\circ u\in {\mathcal C}(T)\) for each y'\(\in Y'\), there exists a unique \(\tilde u\in L({\mathcal C}(T)_ c',E)\) with \(u(t)=\tilde u\circ \delta_ t\) for all \(t\in T\), where \(\delta_ t\) denotes point evaluation at t. The author proves a version of this result without using topology, where \({\mathcal C}(T)\) is a linear subspace of C(T) so that the space of Fourier coefficients of functions in \({\mathcal C}(T)\) is a perfect subspace of \(\ell_ 1({\mathbb{Z}})\) and where E is a perfect linear subspace of \(\ell_ 1\)(\({\mathbb{Z}})\). He applies this to various function spaces on T, e.g. Gevrey classes. However, in all these applications there is a natural l.c. topology at hand.