This is actually a suggestion for the documentation or for a FAQ somewhere.
If I drop a messed up subtitle file (.srt) on the GUI, the charset can’t be changed as the gadget is greyed out. In my example this was a mixture of ISO-8859-1 and UTF-8, but the correct encoding should be ISO-8859-15; it is written in Spanish. Trying to multiplex it “as is” would throw a warning from the GUI and of course had I decided to go ahead the results wouldn’t be pretty.
As soon as I edited the file in an editor and saved it with the correct encoding (it took me several tries, I used MousePad on Linux), the GUI did let me choose whatever charset I wanted for the fixed file and multiplex it.
For text subtitles the GUI will let you change the character set unless the file starts with a BOM. As a BOM signals that a file is encoded in UTF-8, there’s usually no use in letting the user change the character set. In fact, due to mkvmerge always treating the file as UTF-8 it would really be confusing to let the user chose something in such a case.
In short, your file was messed up/broken/wrongly encoded in the first place. As there’s no magic way to fix this, MKVToolNix doesn’t even try. Using another tool that allows for more direct inspection of the results of changing character sets (such as editors) is the preferred way to fix such files.