Linguistic references only; no paper text, tables, code or corpus redistributed.
Samuel Brody and Nicholas Diakopoulos (2011), Cooooooooooooooollllllllllllll!!!!!!!!!!!!!! Using Word Lengthening to Detect Sentiment in Microblogs. EMNLP, pp. 562–570. https://aclanthology.org/D11-1052/
Jacob Eisenstein (2013), What to do about bad language on the internet. NAACL-HLT, pp. 359–369. https://aclanthology.org/N13-1037/
The implemented rule is narrower than general repeated-letter normalization: it only verifies existing lowercase strings with triple vowel/final-letter runs against independently approved English roots, preserving all valid alternatives without copying inflection flags or choosing a canonical spelling by corpus frequency. These forms are labeled derived informal orthographic variants, not individually source-attested entries.
