\documentclass[english]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{CJKutf8}
\usepackage{babel}

% It is impossible to get the document language because the document could start with a command,
% whitespace, and English word or whatever and the rest is in Japanese. Checking for the first CJK
% environment is no solution because the document could be English and contain only some Japanese.

\begin{document}
 
\begin{CJK}{UTF8}{}%
Japanese \end{CJK} \begin{CJK}{UTF8}{}Chinese traditional\end{CJK} \begin{CJK}{UTF8}{}
Japanese \end{CJK}
hello
\begin{CJK}{UTF8}{}%
Chinese simplified \end{CJK}\begin{CJK}{UTF8}{hei} Japanese \end{CJK} \begin{CJK}{UTF8}{}Chinese simplified
\end{CJK}

\begin{CJK}{UTF8}{}
Big5 文鼎楷書
\end{CJK}

\inputencoding{latin9}%
\selectlanguage{english}%
English

\begin{CJK}{UTF8}{}
日本語の文章
\end{CJK}

hello

\begin{CJK}{UTF8}{}%
Korean
\end{CJK}

\end{document}