Well it recognizes a as A and e as E because it's part of the ASCII character set. When a program supports any given character set (ASCII, SJIS, Unicode, UCS, ISO-8859-1, etc.) it must be programmed with the 'knowledge' of what each character 'means' (if it has any specific meaning). Since ASCII was designed in the 60s by Americans it only supports English characters and so it doesn't have any special characters which mean ä or ö or any other variations. Although many fonts now use certain extended characters (ASCII #128-255) of ASCII to display as those characters, they do not have any meaning as letters according to the ASCII character set.
Unicode however has support for practically every character for language and mathematics that's in use today, so when mIRC supports that it should correctly recognize ä as Ä, á as Á, ö as Ö, and so on.
