Rungus and English/Ambiguity Resolution

From LING073
Revision as of 16:48, 15 April 2025 by Gzhang1 (talk | contribs) (→‎Final Evaluation of Ambiguity)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

drg

Initial Evaluation of Ambiguity

  • Parallel Sentence Corpus drg.sentences.txt: ~1.129 (287/254)
  • Largest Monolingual Corpus drg.corpus.basic.txt: ~1.181 (2585/2188)

Identified Ambiguity

Morphological Ambiguity

  • ^Boli/Boli<n>/Boli<v>$ ^po/po<adj>$ ^ka/ka<v>$ ^do/do<det><ind><acc>$ ^vaig/vaig<n>$^,/,<cm>$ ^tuliohon/tuliohon<adj>$ ^oku/oku<prn><pers><p1><sg><nom>$ ^ma/ka<v><iv><impf>$^./.<sent>$
  • ^Olomu/Olomu<adj>$ ^mang/mang<adv>$ ^ino/ino<det><dem><med><nom>$ ^boli/boli<n>/boli<v>$ ^oi/oi<ij>$^./.<sent>$

To resolve this ambiguity, we wrote code that chooses the verb translation when 'boli' when it is sentence initial due to the VSO structure of Rungus.

Lexical Ambiguity

  • ^Popungo<v>/Put<v>$ ^sid<pr>/for<pr>/at<pr>/in<pr>/on<pr>/out<pr>$ ^abai<n>/loft<n>$ ^ino<det><dem><med><nom>/that<det><dem><med><nom>$ ^ikam<n>/mat<n>$ ^dino<det><dem><med><acc>/that<det><dem><med><acc>$ ^nga<cnj><coo>/but<cnj><coo>$ ^omomos<adj>/upright<adj>$^.<sent>/.<sent>$^.<sent>/.<sent>$
  • ^Kada<vaux>/Do not<vaux>$ ^*momozo/*momozo$ ^*inot/*inot$ ^*a/*a$'^*azagan/*azagan$ ^ku<prn><pers><p1><sg><gen>/my<prn><pers><p1><sg><gen>$ ^do<det><ind><acc>/a<det><dem><ind>$ ^tapung<n>/flour<n>$^,<cm>/,<cm>$ ^*pongiduan/*pongiduan$ ^ku<prn><pers><p1><sg><gen>/my<prn><pers><p1><sg><gen>$ ^do<det><ind><acc>/a<det><dem><ind>$ ^dupot<n>/animal<n>$ ^sid<pr>/for<pr>/at<pr>/in<pr>/on<pr>/out<pr>$ ^tapung<n>/flour<n>$ ^ino<det><dem><med><nom>/that<det><dem><med><nom>$^.<sent>/.<sent>$^.<sent>/.<sent>$

Final Evaluation of Ambiguity

  • Parallel Sentence Corpus drg.sentences.txt: 286 / 253 (~1.13043478260869565217)
  • Largest Monolingual Corpus drg.corpus.basic.txt: 2581 / 2184 (~1.18177655677655677656)

Figures are a little strange, likely due to updates made to the lexd file between the initial evaluation and the final evaluation. The top number went down though hehe.