Apertium-tat, a morphological transducer and Constraint Grammar-based tagger for Tatar
1 Rarely Asked Questions
1.1 How do I force an irregular stem to take front/back vowel endings?
For a stem in apertium-tat.tat.lexc file, add {ъ} (i.e. lowercase Cyrillic ‘hard sign’ in curly braces, which have to be escaped in HFST/LEXC) to the right-hand side of the entry if that stem should take back vowel endings, e.g.:
төньяк:төньяк%{ъ%}
көньяк:көньяк%{ъ%}
The lowercase Cyrillic ‘soft sign’ {ь} triggers front vowel harmony:
бәла:бәла%{ь%}
This two signs are only necessary for irregular words when you see that phonological rules from the TWOL file don’t work ‘out-of-the box’ for the word in question.
A major execption are abbreviations. It might be necessary to add either of these two marks to most of them.
СССР:СССР%{ъ%} ABBR ; ! "USSR"