Opening a mailmerge document, Word automativally chooses the wrong language
for the data file (chooses Japanese instead of Windows (ANSI)). The result is
that the merge document can not identify the columns right.

The data is stored in semicolon separated text file, using Norwegian
language (some of the column headings include Norwegian special letters). If
I go and manually choose the data source and then choose Windows (ANSI) as
the data file language, it goes alright (but it is very complicated having to
go through that procedure every time I open the document, so I hope there
should be some way to avoid it).

Thank you in advance for any help!