Wordification: A New Way of Teaching English Spelling Patterns
Paper · 2023 · Lexington Whalen, Nathan Bickel, Shash Comandur et al.
Summary
Wordification is a web application for grade-school children that teaches English spelling by explaining the regularities beneath it instead of relying on memorized word lists. The authors argue that English spelling looks chaotic because letters and sounds do not map one-to-one, but the mismatches follow linguistic logic, and children's spelling errors cluster around those same patterns. The system is built on an 850-word lexicon drawn from BSVL, which covers more than 80% of the words elementary pupils use in their writing. Each entry records phonemes, graphemes, morphology, meaning, grade level, example sentences and audio. Dialect support is a central design goal. Recordings currently exist only for General American English, and AAE and SWE pronunciations are planned, along with careful-versus-casual speech contrasts for ESL learners. The paper describes the technical stack and its migration from Django to Redwood, but reports no learning-outcome evaluation. Its contribution is a design and a linguistically annotated resource. The motivation is that individual tutoring is too costly for most families, while classroom spelling drills consume a lot of time for little gain.
Key points
- The motivating figures are that reading proficiency is reached by just 35% of students in grades 4 and 8, and writing proficiency by 27% of 12th graders. Tutoring costs $50–100/hour, and roughly 60–100 hours are needed to advance a child one grade level.
- Teachers' informal reports suggest current spelling instruction fails about 20% of pupils, is unnecessary for another 20%, and is only marginally useful for the remaining 60%. Spelling takes up about 30 minutes a day, or roughly 10% of class time.
- The authors frame English spelling difficulty through four transparency principles, each of which English breaks:
- one symbol should mean one sound (English breaks this with "c" in cane vs. cent);
- one sound should have one spelling (English breaks this with [k] as c, k or ck);
- one symbol should stand for a single sound (English breaks this with "x" in tax);
- one sound should be written with a single symbol (English breaks this with "ph").
- The lexicon holds 850 word-objects selected and grade-coded according to BSVL. Words are tagged for:
- phonological, orthographic, morphological and semantic properties;
- grade level;
- context of use;
- dialect variation.
- The stack is Next.js, GraphQL, Prisma and TypeScript, with the backend moved from Django (REST) to Redwood. The system supports several user roles, pausable per-student games with progress logging, and a planned link to spell-checkers.
Interesting ideas
- I1. The four transparency principles work as a checklist for diagnosing why a particular letter–sound mapping is irregular. This is more informative than a flat list of exceptions.
- I2. Spelling mistakes are treated as patterned rather than random, because they gather around known grapheme–phoneme mismatches. Error handling can therefore be organized by linguistic cause instead of by surface string.
- I3. Each lexicon entry stores several pronunciations: across GAE, AAE and SWE, and in careful versus casual speech. This lets a learner connect a reduced or dialectal form back to its dictionary form.
- I4. Words are coded with a grade level taken from an established school vocabulary list (BSVL). This gives every item a measurable difficulty.
- I5. The planned loop sends a learner's real typing errors from a spell-checker into the next practice session. That way, training targets exactly what the learner got wrong.
For Mova
How these ideas could serve Mova, our small open language model — its files, stages and tests:
- I1 (tokens): Morphology and spelling-normalization code could tag each non-standard or misspelled form with the principle it trips over, so that every normalization step carries a human-readable reason.
- I2 (learn): Self-induced repair rules could be grouped and named by the underlying mismatch they fix, which makes them easier to audit and to generalize to unseen words.
- I3 (tokens): Lemma lookup could map casual or dialectal spellings onto their standard lemma while keeping a variant tag, so later stages know the original register.
- I4 (answer): Grade tags on vocabulary could be used to calibrate school-course tests and to check whether a student reader is being tested on words appropriate to its level.
- I5 (learn): Failures from each evaluation run could be fed automatically into the next round's training curriculum, focusing silver data and teacher queries on the constructions Mova actually mishandled.
Institutions
United StatesFields, topics and concepts
📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksPsychology · 1,552 worksDevelopmental and Educational Psychology · 499 works
🧭 Topics 1Research subjects and languages the work is about.English · 2,263 works
Cite
Cite as Lexington Whalen, Nathan Bickel, Shash Comandur, Dalton Craven, Stanley Dubinsky and Homayoun Valafar. 2023. Wordification: A New Way of Teaching English Spelling Patterns. arXiv:2309.12981.
@misc{whalen2023wordification,
title = {{Wordification: A New Way of Teaching English Spelling Patterns}},
author = {Lexington Whalen, Nathan Bickel, Shash Comandur, Dalton Craven, Stanley Dubinsky and Homayoun Valafar},
year = {2023},
doi = {10.48550/arxiv.2309.12981},
eprint = {2309.12981},
archivePrefix = {arXiv},
url = {https://doi.org/10.48550/arxiv.2309.12981}
}References
- R. Malatesha Joshi and K. (Kay) Wijekumar, “Introduction: teacher perception, self-efficacy and teacher knowledge relating to literacy,” Annals of Dyslexia, vol. 69, no. 1, pp. 1-4, Jan. 2019, doi: https://doi.org/10.1007/s11881-018-00173-3. “Literacy statistics | United Way of Hays County,” www.unitedwayhaysco.org. https://www.unitedwayhaysco.org/literacy-statistics “The Economic and Social Cost of Illiteracy: A Snapshot of Illiteracy in a Global Context,” Philanthropy News Digest. https://philanthropynewsdigest.org/features/research-briefs/the-economic-and-social-cost-of-illiteracy-a-snapshot-of-illiteracy-in-a-global-context (accessed Mar. 22, 2023). A. C. S. Waldman, “One in Five Americans Struggles to Read. We Want to Understand Why.,” ProPublica. https://www.propublica.org/getinvolved/literacy (accessed Apr. 12, 2023). S. Graham et al., “Teaching Spelling in the Primary Grades: A National Survey of Instructional Practices and Adaptations,” American Educational Research Journal, vol. 45, no. 3, pp. 796-825, 2008, Available: https://www.jstor.org/stable/27667151 D. Kilpatrick, Equipped for reading success: A comprehensive, step by step program for developing phonemic awareness and fluent word recognition. Casey & Kirsch Publishers. Accessed: Apr. 12, 2023. [Online]. Available: https://www.doe.mass.edu/massliteracy/skilled-reading/fluent- word-reading/phonological-awareness.html Horton, R., Johnson, V., & Koonce, N. (2018). From here to there and back again: Forty years of research on African American children’s language. Perspectives of the ASHA Special Interest Groups 3(14): 57-70. Patton-Terry, N., & Connor, C. (2010). African American English and spelling: How do second graders spell dialect-sensitive features of words? Learning Disability Quarterly 33(3): 199-210. Edwards, J. R., & P. Rosin. (2016). A pre-kindergarten curriculum supplement for enhancing Mainstream American English knowledge in non-mainstream American English speakers. Language, Speech, and Hearing Services in Schools 47: 113-122. Johnson, L., Terry, N. P., Connor, C. M., & Thomas-Tate, S. (2017). The effects of dialect awareness instruction on nonmainstream American English speakers. Reading and Writing 30: 2009-2038. Pittman, R. T., Joshi, R. M., & Carreker, S. (2014). Improving the spelling ability among speakers of African American English through explicit instruction. Literacy Research and Instruction 53(2): 107-133. Brown, S. (2011). Listening myths: Applying second language research to classroom teaching. Ann Arbor: University of Michigan Press “Next.js by Vercel - The React Framework,” nextjs.org. https://nextjs.org/ “React,” react.dev. https://react.dev/15.16.17.18. View publication stats Django, “The Web framework for perfectionists with deadlines | Django,” Djangoproject.com, 2019. https://www.djangoproject.com/ IBM, “What is an Application Programming Interface (API) | IBM,” www.ibm.com. https://www.ibm.com/topics/api IBM, “What is a REST API? | IBM,” www.ibm.com. https://www.ibm.com/topics/rest-apis “RedwoodJS Reference,” https://redwood.com/docs/index. https://redwood.com/docs/index 19. 20. Graham, S., Morphy, P., Harris, K., Fink-Chorzempa, B., Saddler, B., Moran, S., & Mason, L. (2008). Teaching spelling in the primary grades: A national survey of instructional practices and adaptations. American Educational Research Journal 45: 796-825. P. M. Cunningham and J. W. Cunningham, “Making Words: Enhancing the Invented Spelling-Decoding Connection,” The Reading Teacher, vol. 46, no. 2, pp. 106– 115, 1992, Available: https://www.jstor.org/stable/20201026 ↗
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/papers/whalen-2023-wordification-new-way-teaching-english/