+
+ DEFVAR_LISP ("word-combining-categories", &Vword_combining_categories /*
+List of pair (cons) of categories to determine word boundary.
+
+Emacs treats a sequence of word constituent characters as a single
+word (i.e. finds no word boundary between them) iff they belongs to
+the same charset. But, exceptions are allowed in the following cases.
+
+(1) The case that characters are in different charsets is controlled
+by the variable `word-combining-categories'.
+
+Emacs finds no word boundary between characters of different charsets
+if they have categories matching some element of this list.
+
+More precisely, if an element of this list is a cons of category CAT1
+and CAT2, and a multibyte character C1 which has CAT1 is followed by
+C2 which has CAT2, there's no word boundary between C1 and C2.
+
+For instance, to tell that ASCII characters and Latin-1 characters can
+form a single word, the element `(?l . ?l)' should be in this list
+because both characters have the category `l' (Latin characters).
+
+(2) The case that character are in the same charset is controlled by
+the variable `word-separating-categories'.
+
+Emacs find a word boundary between characters of the same charset
+if they have categories matching some element of this list.
+
+More precisely, if an element of this list is a cons of category CAT1
+and CAT2, and a multibyte character C1 which has CAT1 is followed by
+C2 which has CAT2, there's a word boundary between C1 and C2.
+
+For instance, to tell that there's a word boundary between Japanese
+Hiragana and Japanese Kanji (both are in the same charset), the
+element `(?H . ?C) should be in this list.
+*/ );
+
+ Vword_combining_categories = Qnil;
+
+ DEFVAR_LISP ("word-separating-categories", &Vword_separating_categories /*
+List of pair (cons) of categories to determine word boundary.
+See the documentation of the variable `word-combining-categories'.
+*/ );
+
+ Vword_separating_categories = Qnil;