Difference between revisions of "Kamba/Disambiguation"

Revision as of 20:03, 9 June 2023

Initial Evaluation of Ambiguity

Link to git: https://github.swarthmore.edu/Ling073-sp23/ling073-kam

Bible Corpus

Ambiguity: ~1.02947229768667160616

Tokens: 24294

Analyses: 25010

Storybook Corpus

Ambiguity: ~~1.02947229768667160616 this number doesn't seem to correspond to the numbers below -JNW

Tokens: 1100

Analyses: 1132

Sentences

Ambiguous words: class 9/10 nouns look the same in the singular and plural.

"Nyũngũ niũkwa"

“The cooking pots are going to break.” -- from Whiteley and Muli (1962) pgs 26 and 139

Nyũngũ nĩũkwa

The cooking pot is going to break. -- our modification to make it singular

Final Evaluation of Ambiguity

Bible

Ambiguity before disambiguation: ~1.02947229768667160616 Ambiguity after disambiguation: ~1.02947229768667160616

Storybook corpus

Ambiguity before disambiguation: ~1.02909090909090909091 Ambiguity after disambiguation: ~1.02909090909090909091

Reasoning

We disambiguated between singulars and plurals of class 9/10 nouns through rules that the noun cannot be singular if followed by a plural verb and vice versa. We know that the rule does work through testing it manually with our own examples as we built the rules, but unfortunately they didn't reduce the ambiguity at all because the specific class 9/10 nouns we have in our lexd file and the few verbs in the few tenses we have so far never appear next to each other in the corpora...