Thanks for the links. I have read them. I will just have to transform
my XML into FO catching every character not in the base font set. I
will then put the character in <fo:inline> and indicate the font
family found in my userconfig.xml file. The bottom line is that the
characters not in the base font set are a wash.

The base fonts are sufficient for most westerners, in particular they
are a slight superset of Latin-1.
If you you have that many non-Latin-1 characters, why don't you use a
full Unicode font, like ArialUni.ttf? You'll have to declare the font
family only once, e.g. on fo:root.

At least I can
display them correctly even if they cannot be searched.

Do you mean you can't find such characters in the resulting PDF?
I don't have this problem.