1: \begin{definition}

2: A probability distribution in $\Delta(T^\infty)$ representing a sequence $X_1,X_2,\ldots$ is said to be \textbf{$(\infty)$-exchangeable} if it is invariant under permutations of finitely many indices\footnote{This finiteness condition is natural in that it allows us to identify $\dsym(T^\infty)$ with the inverse limit $\ilim\dsym(T^n)$ using the Kolmogorov Consistency Theorem ($12.1.2$ in \cite{d:rap}, for example).}, or in other words if the distribution of $X_1,\ldots,X_n$ is $n$-exchangeable for all $n\in\Nm$.  The set of exchangeable distributions is denoted $\dsym(T^\infty)$.  We call a finite or infinite sequence of random variables \textbf{$n$-exchangeable} of \textbf{$(\infty)$-exchangeable}, respectively, if its distribution is.

3: \end{definition}

4: