Elamur Research⌕
☰
home / papers / 1999 / Modeling User Language Proficiency in a Writing Tutor for Deaf Learners of English

Modeling User Language Proficiency in a Writing Tutor for Deaf Learners of English

Paper · 1999 · Lisa N. Michaud and Kathleen F. McCoy

Summary

The paper addresses how a writing tutor can adapt its feedback to learners whose English proficiency varies widely. The target users are deaf students whose first language is ASL, so the authors treat written English as a second language for them. The tutor is ICICLE. Its grammar is extended with error rules that flag mistakes in multi-sentence essays. The authors propose a user model, SLALOM, built on interlanguage theory, the Zone of Proximal Development (ZPD), and the cognitive, associative and autonomous stages of skill learning. SLALOM is an overlay on a knowledge base of morphosyntactic features. The features are arranged in hierarchies that follow orders of acquisition, and lateral links join features that tend to be learned at the same time. Each feature is tagged as acquired, in the ZPD, or untagged if it does not yet appear in the learner's writing. The model would help choose among competing error interpretations, filter which errors are worth teaching, and shape the generated explanations. This is a design proposal. The authors cite only preliminary empirical support and argue that the model gives fine-grained, robust information for language assessment.

Key points

  • ICICLE (Interactive Computer Identification and Correction of Language Errors) analyzes essays with an English grammar extended by rules for errors typical of ASL-native writers. Errors are color-coded by class; for example, subject-verb agreement errors are shown in blue.
  • Three system components consult the model:
    • error identification, to decide whether an error reflects ignorance of a rule, a misapplied rule, or a typo (example from their corpus: "My brother like to go");
    • an error filter between analysis and response, which drops errors outside the learner's current level;
    • the text planner and surface generator, which need to know which concepts and constructs the learner already understands.
  • The model's features have the same granularity as the parser's grammar features. A feature used correctly and consistently counts as acquired, one with variable or erroneous use counts as ZPD, and one that is rare or absent is beyond the learner's level.
  • Model updates come from the error identification module. The model tracks how often each feature is attempted and how often it is executed without error, keeping counts for the latest essay separate from counts for earlier ones. It uses a sliding window over past writing samples, and the right window size is left open.
  • The very first session must run without a user model. It relies instead on expected co-occurrences of errors.

Interesting ideas

  • I1. Each produced sentence yields evidence on many constructs at once, which makes it far richer than a single diagnostic question; free text is a dense assessment signal.
  • I2. Ambiguous errors are disambiguated by asking what the producer probably knows: the same surface mistake can reflect a missing rule, a misapplied rule, or a slip.
  • I3. The model separates three zones: features already mastered, features under active development, and features not yet attempted. Only the middle zone is worth correcting or teaching.
  • I4. Linear acquisition orders are combined with lateral links between co-learned features, so evidence on one feature propagates to its neighbors until direct data contradicts it.
  • I5. Statistics are computed over a sliding window of recent samples, so old performance does not mask current ability and the effect of recent interventions can be judged.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (learn): Score every construction a parser handles in each sentence, rather than one aggregate accuracy, to build a per-feature competence map of Mova's own components.
  • I2 (tree): When several witness parsers disagree, a prior on which constructions each parser reliably handles could decide which analysis to trust, and the reason could be logged explainably.
  • I3 (learn): Self-induced repair rules and training data could target constructions where Mova is inconsistent, skipping both solid cases and phenomena far beyond its current ability.
  • I4 (world): The knowledge graph could support default inference along ordered and sibling links, marked as defeasible and overridden by explicit evidence in the text.
  • I5 (learn): Evaluation could compare recent-window metrics against older runs to check whether a new repair rule or teacher tier actually changed behavior.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of DelawareNewark · United States20 works · 1,056 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.English · 2,263 works

Cite

Cite as Lisa N. Michaud and Kathleen F. McCoy. 1999. Modeling User Language Proficiency in a Writing Tutor for Deaf Learners of English. Computer Mediated Language Assessment and Evaluation in Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{michaud1999modeling,
  title = {{Modeling User Language Proficiency in a Writing Tutor for Deaf Learners of English}},
  author = {Lisa N. Michaud and Kathleen F. McCoy},
  booktitle = {Computer Mediated Language Assessment and Evaluation in Natural Language Processing},
  year = {1999},
  doi = {10.3115/1598834.1598844},
  url = {https://doi.org/10.3115/1598834.1598844}
}

References

26 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. N. Bailey, C. Madden, and S. D. Krashen. 1974. Is there a 'natural sequence' in adult second language learning? Language Learning, 24(2):235-243. find
  2. C. Baker and D. Cokely. 1980. American Sign Language: A Teacher's Resource Text on Grammar and Culture. TJ Publishers, Silver Spring, MD. find
  3. Giuseppe Carenini and Johanna D. Moore. 1993. Generating explanations in context. In Wayne D. Gray, William E. Hefley, and Dianne Murray, editors, Proceedings of the International Workshop on Intelligent User Interlaces, pages 175-182, Orlando, Florida, January 4-7. ACM Press. find
  4. Michelene T. H. Chi, Robert Glaser, and Ernest Rees. 1981. Expertise in problem solving. In Advances in the Psychology off Human Intelligence, chapter 1, pages 7-76. Lawrence Erlbaum, Hillsdale, NJ.
  5. Vivian Cook. 1993. Linguistics and Second Language Acquisition. Macmillan Press Ltd, London. ↗
  6. Heidi C. Dulay and Marina K. Butt. 1974. Errors and strategies in child second language acquisition. TESOL Quarterly, 8(2):129-136, June. ↗
  7. Heidi C. Dulay and Marina K. Burt. 1975. Natural sequences in child second language acquisition. Language Learning, 24(1). find
  8. Rod Ellis. 1994. The Study of Second Language Acquisition. Oxford University Press, Oxford. find
  9. David Ingrain. 1989. First Language Acquisition: Method, Description, and Explanation. Cambridge University Press, Cambridge and New York. find
  10. Stephen Krashen. 1981. Second Language Acquisition and Second Language Learning. Pergamon Press, Oxford. find
  11. Kathleen F. McCoy and Lisa N. Masterman (Michaud). 1997. A tutor for teaching English as a second language for deaf users of American Sign Language. In Proceedings of Natural Language Processing .for Communication Aids, an A CL/EA CL97 Workshop, pages 160-164, Madrid, Spain, July. find
  12. Kathleen F. McCoy, Christopher A. Pennington, and Linda Z. Suri. 1996. English error correction: A syntactic user model based on principled mal-rule scoring. In Proceedings of the Fifth International Conference on User Modeling, pages 59-66, Kailua-Kona, Hawaii, January 2-5. UM96, User Modeling, Inc. find
  13. Lisa N. Michaud and Kathleen F. McCoy. 1998. Planning text in a system for teaching English as a second language to deaf learners. In Proceedings of Integrating Artificial Intelligence and Assistive Technology, an A A A I '98 Workshop, Madison, Wisconsin, July. find
  14. Johanna D. Moore and Cecile L. Paris. 1992. Planning text for advisory dialogues: Capturing intentional and rhetorical information. Computational Linguistics, 19(4):651-695. find
  15. J. Michael O'Malley and Anna Uhl Chamot. 1990. Learning Strategies in Second Language Acquisition. Cambridge University Press, Cambridge. ↗
  16. Cecile Laurence Paris. 1987. The Use of Explicit User Models in Text Generation. Ph.D. thesis, Columbia University. find
  17. Martin H. Ringle and Bertram C. Bruce. 1981. Conversation failure. In W. G. Lehnert and M. H. Ringle, editors, Knowledge Representation and Natural Language Processing, chapter 7, pages 203-221. Lawrence Erlbaum Associates, Hillsdale, New Jersey. find
  18. Robert Rueda. 1990. Assisted performance in writing instruction with learning-disabled students. In Luis C. Moll, editor, Vygotsky and Education: Instructional Implications and Applications of Sociohistorical Psychology, pages 403-426. CUP, Cambridge. find
  19. David Schneider and Kathleen F. McCoy. 1998. Recognizing syntactic errors in the writing of second language learners. In Proceedings of the Thirty-Sixth Annual Meeting of the Association for Computational Linguistics and the Seventeenth International Conference on Computational Linguistics, volume 2, pages 1198-1204, Universite de Montreal, Montreal, Quebec, Canada, August 10-14. COLING-ACL, Morgan Kaufmann Publishers. find
  20. Bonnie D. Schwartz and Rex A. Sprouse. 1996. L2 cognitive states and the Full Transfer/Full Access model. Second Language Research, 12(1):40-72. ↗
  21. Bonnie D. Schwartz. 1998. On two hypotheses of "transfer" in L2A: Minimal Trees and Absolute L1 Influence. In S. Flynn, G. Martohardjono, and W. O'Neil, editors, The Generative Study of Second Language Acquisition, chapter 3, pages 35-59. Lawrence Erlbaum, Mahwah, NJ. find
  22. L. Selinker. 1972. Interlanguage. International Review of Applied Linguistics, 10:209-231. find
  23. Linda Z. Suri and Kathleen F. McCoy. 1993. A methodology for developing an error taxonomy for a computer assisted language learning tool for second language learners. Technical Repor t TR=93-16, Dept. of CIS, University of Delaware. find
  24. Lev Semenovich Vygotsky. 1986. Thought and Language. MIT Press, Cambridge, MA. find
  25. Wolfgang Wahlster and Alfred Kobsa. 1986. Dialog-based user models. In Giacomo Ferrari, editor, Proceedings of the IEEE, Special Issue on Natural Language Processing, February. find
  26. Beverly P. Woolf. 1984. Context dependent planning in a machine tutor. Technical Report COINS Technical Report 84-21, Dept. of Computer and Information Science, University of Massachusetts at Amherst, May. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/michaud-1999-modeling-user-language-proficiency-writing/