Sure, and then we'd be hearing about complaints of how stupid the algorithm for detecting the right language is.
Well, it cannot fail too much.
Check with language 1, calculate the number of mispellings, check with language 2, calculate the number of mispellings, etc...
And pick the one with the lowest ^_^
The worst that can happen is a false negative with some errors going undetected cause they looked like another language.