Northern Haida Disambiguation

This file gives an nverview of some still ad hoc solutions for disambiguation.


  • vislcg3 installed A text corpus. The one we have, test/data/haida_grammar.utf8corpus.txt, is fetched from Jordan's brochure. Only full sentences are included, they are marked with a sentence initial "# " symbol. They all have a translation in English, marked with an initial "&".

Making missinglist

cat test/data/haida_grammar.utf8corpus.txt |grep "#"|cut -c3-|\
tr ' ' '\n' | tr '=' ' '|uhdn|grep '?' |cut -f1 | sort |\ 
uniq -c | sort -nr > doc/data/missing.haidagrammarcorpus

How to analyse

Haida differs from the other languuages in not having an adjusted version of the preprocessor yet. While waiting , we do some ad hoc solutions. Here is a pipeline that gives an analysis.

Note the = symbol, it is used to mark MWEs in the text.

cat test/data/haida_grammar.utf8corpus.txt |grep "#"|cut -c3-|tr ' ' '\n' | tr '=' ' '|uhdn|lookup2cg | vislcg3 -g src/syntax/disambiguation.cg3

Strategies for disambiguation

Look at common ambiguity patterns

To create similar statics, use the script (write --help in order to get just that.