All annotations should correspond to the lemma

Many annotations for SemCor do not correspond to lemmas in English Wordnet. There seem to be a few reasons for this

1. Many proper nouns are tagged with a broader category. The first annotation in the corpus is 'Fulton County Grand Jury' as a "group", instead we can use the Namenet targets and tag it to Wikidata and Wordnet, with `Q486633` for 'Fulton County' and `08431921-n` for 'grand jury'
2. Sometimes the lemmas don't align because of split verbs or particles that don't occur contiguously with the text
3. Open English Wordnet has been updated in a way that means that we must reannotate.

Closing this issue also requires that the validation script is updated to verify that all lemmas match as in English Wordnet.

Start by reviewing the SemCor annotations, the Namenet targets, and the current validation script, then compare their lemmas with the relevant English Wordnet release. Done means the annotations cover proper nouns and split verbs or particles consistently, remain aligned with Open English Wordnet, and validation verifies every lemma match.

Written by the indexing model from the issue text.