Back to LanguageTool Homepage - Privacy - Imprint

How add additional grammar information to POS dictionary?

(Yakov) #1

How add additional grammar information like valency, etc. to POS tag dictionary?

(Tiago F. Santos) #2

There are more qualified persons than me to reply this, but I hope I can help.

I can see two ways of achieving that.

One is rebuilding the dictionary and adding that information to existing tags.

As an example, you may replace in the extracted list, instances of VB:Past:Fem with VB:Past:Fem:Val1. This may be laborious and requires adaptation of many existing rules.

Another way, I have seen in some language projects used here, is to create a pseudo-chunker with disambiguation.xml. You can add the POS tag attributes you want to given words.
For example, you can add a rule that recognizes all inflected forms of X, Y and Z verbs and adds the POS tag 'VAL_1' (you can also do this for verb arguments, though that is as time consuming as developing that part of a chunker). You can use those tags in your new rules. The pros of this method is that you do not have to rebuild rules, the cons is that the rules will be more confusing to write and understand.