I am encountering this same problem: trying to add a word that
contains a special character into a personal dictionary gives an error
like:
@(#) International Ispell Version 3.1.20 (but really Aspell 0.60.3)
Error: The word "erf?rt" is invalid. The character '?' (U+3F) may not
appear in the middle of a word.
The word is utf-8-encoded into a file that is piped to the Aspell
command. The command includes the --encoding='utf-8' option. The
personal dictionary file also specifies utf-8 as encoding, and if I
add the word manually to this dictionary, it is correctly taken into
account when spell checking is done.
Apparently, Aspell ignores the option specifying the encoding of the
input, and assumes that it is encoded in the same character set as
the main dictionary.
I don't think that is the issue. The issue is that the Aspell is 8-bit
internally and the personal dictionary needs to get mapped to the same
character set that the main dictionary uses. The '?' means that the
Unicode character was not available in the main dictionary character set
and it thus defaulted to a '?'.
But, if I am using a dictionary for Spanish and try to add to a
personal dictionary a word that contains the character รณ, which is
very common in Spanish, I get this same message. Obviously it is very
simple to convert this character from utf-8 to iso-8859-1, the latter
being apparently used by the Spanish dictionary...