French educator building evidence-informed language-learning infrastructure.

I teach French and build Concordance, a language-learning platform for serious, intentional learners and teachers. The project is grounded in demonstrated communication, learner agency, teacher judgment, cultural connection, privacy, and meaningful proficiency rather than completion metrics or engagement loops.

Concordance is an evidence-informed language-learning platform designed to help people become more capable and independent communicators. French is the first implementation, but the longer-term goal is a language-neutral foundation that can support learners across languages, educational settings, and professional contexts.

The project does not assume there is one ideal curriculum or machine-predicted path through a language. Learners have different goals and their own pace. Concordance is being built to identify meaningful opportunities from evidence while preserving learner choice and keeping teacher-required work authoritative.

Language and culture are inseparable. Over time, I want Concordance to draw on different kinds of linguistic and cultural expertise, including educators, linguists, translators, heritage speakers, community members, and younger speakers connected to contemporary usage. Any future work involving endangered or Indigenous languages must be undertaken with the communities involved, not simply about them.

LAOS is the evidence-centered educational reasoning architecture beneath Concordance. At a deliberately high level, its progression is:

Evidence
→ Evidence Assertions
→ Learner State
→ Framework Interpretation
→ Recommendations
→ Individual Learning Opportunities

The distinctions are intentional. An observation is not automatically a conclusion about proficiency. A proficiency interpretation is not automatically a recommendation. A recommendation is not a command. When evidence is insufficient, the system should be able to preserve that uncertainty rather than manufacture an answer.

Substantial parts of this architecture now exist. The foundations for evidence assertions, canonical learner state, framework projection, recommendation, and the Individual Learner Path have moved from design into implementation. Current work continues to close the ILP architecture and connect this educational intelligence to useful learner experiences without allowing delivery mechanics to redefine educational truth.

ACTFL and CEFR can provide independent perspectives on demonstrated proficiency without becoming the learner record itself. AI can assist with bounded tasks, but it does not become the educational authority. Teachers retain professional judgment and learners retain meaningful agency over their learning.

As Concordance becomes more capable of understanding learner evidence, privacy becomes more important, not less.

A principle guiding the architecture is:

The fact that Concordance possesses information does not mean every part of Concordance is entitled to use it.

Writing, speech, recordings, mistakes, and patterns of development can reveal a great deal about a learner. Educational evidence should not quietly become behavioral surveillance, and more data is not automatically better education. This responsibility is especially important when building technology that may be used by children and adolescents.

A standalone tool for generating contextualized vocabulary gap-fill activities.

• National Board Certified Teacher — Early Adolescent/Young Adult World Languages, French
• AP French Language and Culture exam Reader
• IB examiner, Language B French
• Treasurer, Arizona AATF
• Active member, Alliance Française of Greater Phoenix
• M.A. French Language and Literature, University of Pittsburgh
• M.Ed. Secondary Education, Grand Canyon University
• B.A. French, Summa Cum Laude, Arizona State University

My work combines classroom teaching, applied language pedagogy, Francophone literature and culture, educational technology, and responsible AI.

• Evidence before recommendation
• Teachers retain educational judgment
• Learners retain agency, choice, and their own pace
• Communication matters more than streaks or screen time
• AI assists educational work but does not own educational truth
• ACTFL and CEFR are useful perspectives, not learner identities
• Language learning should remain connected to living cultures and communities
• Privacy, accessibility, and affordability are part of responsible design

I build Concordance while teaching full time and raising a young family. My goal is to keep serious language-learning tools broadly accessible while making the project sustainable enough to support continued development, infrastructure, and eventually the educators, language specialists, and community contributors whose expertise can help it grow.

Individual sponsorship, institutional partnerships, and research support can help make that possible without giving funders authority over educational conclusions or learner data.

If the project is useful to you or its mission resonates with you, consider sponsoring development on GitHub or following the public Concordance repository.

• Concordance: concordancelearn.com
• Public project: github.com/monsieur-trenton/concordance
• LinkedIn: linkedin.com/in/john-trenton