Process convergence of self-normalized sums of i.i.d. random variables coming from domain of attraction of stable distributions (Q1948993)

Let \(X_i\) be i.i.d. symmetric observations from the domain of attracion on an \(\alpha\)-stable distribution. Let \[ S_n(t)= \sum^{[nt]}_{i=1} X_i\quad\text{and}\quad V_{n,p}= \Biggl(\sum^n_{i=1} |X_i|^p\Biggr)^{{1\over p}}\;(p> 0) \] and define the process \(Y_{n,p}\) by \[ Y_{n,p}(t)= {S_n(t)\over V_{n,p}}+ (nt- [nt]) {X_{[nt]+ 1}\over V_{n,p}}\quad (0< t< 1). \] It is known that Donsker's theorem for \(Y_{n,2}\) holds if and only if \(\alpha 2\) (cf. [\textit{M. Csörgő}, \textit{B. Szyszkowicz} and \textit{Q. Wang}, Ann. Probab. 31, No. 3, 1228--1240 (2003; Zbl 1045.60020)]). In this paper, to obtain the limit distribution of \(Y_{n,p}\), the authors consider the four cases (i) \(0< p<\alpha\leq 2\), (ii) \(0< p=\alpha< 2\), (iii) \(p>\alpha\), (iv) \(p= \alpha= 2\), and show that, in the cases (i) and (ii), the finite-dimensional distributions converge in probability to a degenerate distribution at zero and in the cases (iii) and (iv) finite nontrivial limiting distributions exist, but (iv) is the only case in which both tightness and finite-dimensional convergence hold, and hence \(Y_{n,p}\) converges weakly to Brownian motion in \(C[0,1]\).