Misapplied exceptional forms? #43
Description
This is for the conversion of Princeton WordNet files for the English omw-en* wordnets.
As noted in this thread goodmami/wn#199 (comment), the exceptional forms (in noun.exc, verb.exc, etc.) are perhaps over-applied. The issue is that the lemma form was matched against case-normalized lemmas, leading to situations like this:
noun.exc
buffaloes buffalo
omw-en31.xml
<!-- This entry is for the city of Buffalo -->
<LexicalEntry id="omw-wn31-Buffalo-n">
<Lemma writtenForm="Buffalo" partOfSpeech="n" />
<Form writtenForm="buffaloes" />
<Sense id="omw-wn31-Buffalo-09141172-n" synset="omw-wn31-09141172-n" dc:identifier="buffalo%1:15:00::" />
</LexicalEntry>
<!-- This entry is for the meat and the animal -->
<LexicalEntry id="omw-wn31-buffalo-n">
<Lemma writtenForm="buffalo" partOfSpeech="n" />
<Form writtenForm="buffaloes" />
<Sense id="omw-wn31-buffalo-02413348-n" synset="omw-wn31-02413348-n" dc:identifier="buffalo%1:05:02::">
<Count>3</Count>
</Sense>
<Sense id="omw-wn31-buffalo-07679237-n" synset="omw-wn31-07679237-n" dc:identifier="buffalo%1:13:00::" />
<Sense id="omw-wn31-buffalo-02410605-n" synset="omw-wn31-02410605-n" dc:identifier="buffalo%1:05:01::" />
</LexicalEntry>My guess is we don't actually want the exceptional plural for the proper name entry. So I propose doing a case-sensitive match for lemmas of the exceptional forms.