The longest words of the corpus with minimum frequency 2 are shown. The words are seen at least twice, hence, there is some chance for not seeing misprinted words.
Surprisingly, there is no longest word which is much longer than the second one. This, again, argues for correct preprocessing.
In the case of correct preprocessing, the longest words are true words. In many cases, they belong to some topics which can generate these long words.
In the case of poor preprocessing, some non-word strings will appear.
The length of the longest words clearly depends on language and corpus size.
select char_length(word) as le, freq, word from words where freq>1 order by le desc limit 10;
How does the length of the longest words increase with corpus size?
3.2.3.1 Longest Words in top-1000 by length