Word is a sequence of characters.
LF or CR or LFCR are the interrupt characters to change paragraph.
When inside word have a character from a different language or random Unicode character, except LF, CR, LFCR and SPACE, in Scribus 1.4.2 word breaks in two, but in Scribus 1.3.3.12 not, that is correct.
How can fix it?