If you're doing much copy-paste from word processors, then it's better to load
in your LaTeX document. Then you can teach LaTeX new correspondences between Unicode characters and LaTeX set of known objects, in this case
%% the first is the “unknown minus” (U+2212), the second is a hyphen
How do you know there's something amiss? Look in the .log file: you should find several instances of something like
Missing character: There is no <E2> in font cmr10!
Missing character: There is no <88> in font cmr10!
Missing character: There is no <92> in font cmr10!
one for each “disappearing” minus sign. Indeed, if you don't load
inputenc with the
utf8 option, TeX just sees the three byte representation of U+2212, precisely the bytes
0x92 that don't correspond to a printable character, because the default TeX fonts have only 128 glyphs.
utf8, but no
\newunicodechar declaration you'd get an error:
! Package inputenc Error: Unicode char \u8:− not set up for use with LaTeX.
and you can so understand what went wrong: the minus sign is not understood. Copy it and use it for defining the correct