Last updated
How MorphyLearn is built — and what we have tested so far
MorphyLearn is a new, family-built literacy platform focused on helping children explore how longer words are built from meaningful parts. This page explains how the curriculum is created, how missions make decisions, how we internally test the software, and where the evidence currently stops.
We're early — so we're showing our work
MorphyLearn is a new, family-built literacy platform. We do not yet have years of classroom use, an independent efficacy study, or a controlled comparison involving real learners. We will not present internal curriculum checks or software simulations as if they were learner-outcomes evidence.
What we can show today is how the curriculum is constructed, how words and examples are checked, how the mission engine makes decisions, and how we test that the software follows those rules.
As MorphyLearn grows, outside curriculum review, accessibility testing, real-world pilot data, and independent learner-outcomes research are additional layers we hope to pursue and report separately.
Four different questions require four different kinds of evidence
These layers should not be treated as interchangeable. Each section on this page is labeled with which one it belongs to.
Why teach word parts?
Morphological awareness — recognizing and working with meaningful parts such as roots, prefixes, and suffixes — is an established area of literacy instruction. This supports the general instructional approach, not MorphyLearn's effectiveness.
Are the words and examples sound?
MorphyLearn internally checks word structure, lexical evidence, definitions, examples, and content standards before content is introduced or revised. A newer formal whole-curriculum sign-off process is still ongoing.
Does the software follow its rules?
Automated tests and synthetic-learner simulations check mission selection, scoring, review scheduling, progression rules, achievements, and data separation.
Does it improve outcomes for real children?
MorphyLearn has not yet completed an independent study showing gains in reading, vocabulary, spelling, retention, comprehension, or transfer compared with another form of instruction.
Morphology is an established part of literacy instruction
Morphological awareness means noticing and working with meaningful parts of words. Roots, prefixes, and suffixes can give learners useful clues about word structure and meaning. MorphyLearn is one implementation of this broader instructional idea — not the source of the idea itself.
- Morphological Awareness: One Piece of the Literacy Pie (opens in a new tab) International Dyslexia Association
- Morphology 101 (opens in a new tab) International Dyslexia Association
These are outside literacy resources about morphology. Their inclusion does not mean that the International Dyslexia Association has reviewed, certified, partnered with, or endorsed MorphyLearn.
Written first. Checked before use.
Words, definitions, examples, and answer choices are prepared before they appear in a child's mission. MorphyLearn does not generate new curriculum text live while a child plays. Content is written and reviewed internally by MorphyLearn's creators using the 20-dimension standard below — an internal authoring and review standard, not a third-party certification. Anything that fails a hard rule is revised or dropped, not shipped anyway.
See MorphyLearn's internal 20-dimension content and bias review
Language and meaning
- Stays true to the word's real dictionary meaning
- Gives enough context to actually infer the word's meaning
- Reads like something a real person would say
- Reads as an ordinary sentence, not a lesson
Instructional usefulness
- Matches the vocabulary and complexity of the right age level
- Family references feel natural, never forced
Emotional safety
- No trauma, abuse, or frightening conflict
- Feelings are shown neutrally, never as a flaw
- No "lazy," "bad," or similar labels applied to a person
- Never references a real family's actual private life — health, discipline, or real conflict, even indirectly
Representation and agency
- Doesn't default any role to one gender
- Interests and emotions aren't tied to gender
- Caring for others is mutual, never an assigned duty
- Protecting someone never means controlling their choices
- Leadership is shared or earned, not defaulted to one role
- Kid characters get to make real choices, not just react
Assessment fairness
- Tests real understanding, not just recall
- Wrong answers are ordinary and fair, never a trick
- The right answer never just restates the word's own definition
Whole-curriculum pattern checks
- Checked across the whole curriculum, not just one sentence at a time, so no single casting pattern quietly dominates
Live curriculum count as of August 7, 2026. The curriculum keeps growing through this same write-ahead, internally-reviewed process — these numbers are the actual count running in production, not a fixed or rounded figure.
A word must be more than mechanically buildable
Real prefixes, roots, and suffixes can sometimes be combined into something that looks plausible without being an established English word. MorphyLearn does not treat every possible combination as valid — candidate words are checked across multiple external lexical sources and against their stored word-part structure before they're used with children.
Morphology reconstruction. Does the stored prefix, root, and suffix structure reconstruct the intended word correctly?
Lexical attestation. Is the word supported across multiple external lexical sources — WordNet (a lexical database of words, senses, and parts of speech), wordfreq (corpus-frequency evidence, not a dictionary), and spylls/Hunspell (spelling-dictionary and productive-affix evidence)?
Frequency and rarity. A rare word is not automatically rejected. Low-frequency words receive a distinct status so rarity is visible rather than mistaken for invalidity.
Definition and example review. Part-of-speech analysis and custom mechanics checks help surface missing target words, wrong grammatical roles, circular definitions, duplicate examples, and content that needs human interpretation — an automated tagger is a signal, not the final word.
Human decisions. Automated evidence can surface uncertainty, but it does not silently make every semantic judgment. Items needing interpretation remain in a review queue rather than being auto-approved or auto-rejected.
Definitions are written and reviewed during curriculum authoring. MorphyLearn's newer verification system deliberately does not grant automatic semantic approval based only on dictionary matching — formal sense-by-sense sign-off remains a separate human review process, distinct from the authoring review every definition already goes through.
MorphyLearn's full-curriculum formal verification process is still ongoing. Stable, completed findings are published here as their scope becomes clear — for example, a printable deck isn't labeled ready until every word and its required content pass that deck's own release gate. This page will be updated as the whole-curriculum review progresses.
How a mission gets built for your kid, in order
Regular Word Lab missions are not pre-scripted. Each time your child opens a regular mission, the app decides what to show at that moment, in this order:
Start with what they haven't solidly learned yet. Words at their current level, not already-mastered ones, come first — round-robined across word families so two words from the same family are never placed back to back, as long as another family still has words left to interleave with.
Blend in a small stretch, once they've earned it. After a child has learned at least 10 words at their level, with 60% of them solid, missions start mixing in a few harder words — capped at 3 in every 10, never a flood.
Roughly every 4th mission is a Memory Mission. A shorter review round that resurfaces words they've missed, words they haven't seen in a while, and words worth double-checking still stuck.
Sometimes, a Mystery Word. Once your child has used a specific prefix or suffix correctly at least 5 times across at least 3 different words, with at least 85% accuracy, a Memory Mission may include a new curriculum word built from that same part — one they've never practiced before. This looks for an in-app signal that the learner can apply a familiar part to an unfamiliar word.
MorphyLearn treats this as an internal transfer signal. It has not yet been validated as a research measure of transfer to reading outside the app. One correct answer isn't proof of broad transfer.
Built to never feel like a test
No fail marks
MorphyLearn is designed to avoid marking a wrong answer as a failure. A wrong first try just prompts another look. A wrong second try is shown as "now you know it" with a graduation cap — not an X, not a red mark.
No timers, ever
MorphyLearn is designed to avoid any countdown. Response time is never used to score, rank, or pressure a kid into rushing.
No streak-loss guilt
MorphyLearn is designed to avoid loss messaging around streaks. A streak is shown when your child has one — positively, and only then — with no broken-streak notification.
How progress is tracked
Your parent dashboard shows two separate signals, deliberately not blended into one score because they answer different questions. Both describe in-app behavior, not an outside assessment.
"Mastered" badge
A simple, dashboard-facing signal: your child got the word right on the first try at least once, and it isn't currently flagged for review. "Mastered" is MorphyLearn's dashboard label for this specific in-app rule — it is not a standardized, clinical, or independently validated measure of mastery.
"Automatic" status
The stronger signal that actually changes what your child sees: the child answered the word correctly on the first try three times in a row, and answered correctly again during a separate Memory Mission — showing the word was answered correctly again later within MorphyLearn. Only then does it stop appearing in normal rotation.
These two can genuinely disagree for the same word at the same time — a word can be "Automatic" (no longer needing regular practice) while the simpler badge hasn't caught up yet. That's by design, not a bug: one is a quick dashboard read, the other is what the engine itself uses to decide what to show next.
How we test the learning engine
Anyone can say software is well tested. Here is what MorphyLearn's internal validation actually checks — and what it does not.
What we actually did
Built four synthetic learners with different play styles — one guessing at random, one improving steadily, one performing well from the start, and one who struggles but keeps playing every mission anyway. Real kids don't all learn the same way, so one test profile wasn't enough.
Played real missions using the real engine. Not a mock, not a simplified stand-in — the same session, answer-checking, and progress-tracking code path a real browser calls when your child plays.
Ran it against a temporary copy of the production curriculum database — a temporary copy of the 1,033-word production curriculum as it existed during the August 1 test run, not a mock or reduced test set. Safely: the live database is opened read-only, a full private copy is made, testing runs against that copy only, and the copy is deleted immediately after. The real database is never written to, and the synthetic test accounts are structurally separate from real family accounts and never appear in anyone's real reports.
Checked tens of thousands of automated rule assertions after every single mission — did the points awarded match what was actually earned, did a badge only flip on when it should have, did nothing leak into a real household's records. Not "did it run without crashing" — "did every single outcome match what should have happened."
What came back
| What was checked | Result |
|---|---|
| Simulated missions completed | 2,400, across four synthetic learner profiles |
| Curriculum coverage | All 1,033 active words were reached by each of the four synthetic learner profiles during the reported run |
| Engine crashes / errors | 0, across all 2,400 simulated missions |
| Automated rule assertions | 55,020 run, 55,020 passed |
| Achievement badges | Verified to flip on at the exact right moment and stay on afterward |
The one result that wasn't a clean sweep is the most informative part: the "struggles but keeps playing" learner never earned the badge for a mistake-free mission, across all 300 of its simulated missions. That badge specifically requires a mission with zero wrong answers — a learner modeled to make mistakes throughout correctly never qualified for it, while every other, less demanding badge fired normally for that same learner. This shows the badge's rule was implemented consistently, not that the achievement system is educationally valid.
In this internal simulation, the tested version of the engine behaved consistently with its programmed rules across the reported scenarios.
These results do not show that children gain vocabulary, improve reading or spelling, retain learning over time, or transfer skills outside MorphyLearn. They are not an independent audit, classroom trial, clinical study, or comparative outcomes study.
Where the evidence stands today
| Evidence layer | Current status |
|---|---|
| Morphology as a literacy concept | Supported by outside research |
| MorphyLearn word/curriculum checks | Internal, in progress |
| MorphyLearn software behavior | Internally tested |
| Real-world use | Early use; no formal usability study completed |
| Learning outcomes | Not yet established |
| Comparative effectiveness | Not established |
| Independent curriculum review | Not yet completed |
| Independent accessibility audit | Not yet completed |
Questions about how this all works
Does the simulation prove that children learn from MorphyLearn?
No. It tests whether the software follows its programmed rules — session logic, scoring, badges, and data separation. Measuring learning outcomes requires real learners, appropriate assessments, and ideally an independent study design or review. MorphyLearn has not yet completed that kind of study.
Has MorphyLearn been independently reviewed?
Not yet in the sense of an outside curriculum audit, accessibility audit, or learner-outcomes study. Everything described on this page — the content rubric, the lexical checks, the software simulations — is MorphyLearn's own internal work. Linking to outside resources about morphology, further up this page, is not a review or endorsement of MorphyLearn by those organizations.
Are all curriculum words checked in dictionaries?
Not quite — "checked in dictionaries" understates it. MorphyLearn checks candidate words against multiple external lexical sources that provide different kinds of evidence: WordNet (a lexical database), wordfreq (corpus-frequency evidence), and spylls/Hunspell (spelling-dictionary and productive-affix evidence) — plus MorphyLearn's own morphology-reconstruction check and, for definitions and examples, part-of-speech analysis. A word missing from one source isn't automatically rejected; English word formation is productive enough that real, rare words sometimes are.
Does MorphyLearn replace phonics or reading intervention?
No. MorphyLearn teaches morphological awareness — recognizing and working with meaningful word parts — which is one piece of literacy instruction, not a replacement for phonics instruction or a structured reading intervention a struggling reader may need.
What does "Mastered" or "Automatic" mean?
"Mastered" is MorphyLearn's dashboard label for a specific in-app rule: the child got the word right on the first try at least once, and it isn't currently flagged for review. It is not a standardized, clinical, or independently validated measure of mastery. "Automatic" is a stronger in-app signal: the child answered correctly on the first try three times in a row, and answered correctly again later during a separate Memory Mission. Both describe in-app behavior, not an outside assessment.
Has transfer been proven?
No. A Memory Mission may occasionally include a new curriculum word built from a word part the child has practiced — an in-app signal that the learner can apply a familiar part to an unfamiliar word. MorphyLearn treats this as an internal transfer signal, not as a validated research measure of transfer to reading outside the app. One correct answer is not proof of broad transfer.
Is "the engine has been tested" just a marketing line?
No — it's from a real, dated run: 2,400 simulated missions played by the actual engine code against a temporary copy of the production database, checked against 55,020 automated rule assertions, with 0 failures, on August 1, 2026. The full methodology, including the one badge that correctly never fired for a struggling learner, is laid out above — along with what this testing does and does not establish.
Does testing the engine use my child's real data?
No. Testing runs against a temporary, private copy of the production curriculum database that's created and automatically deleted for each test run — the live database itself is never written to. The synthetic accounts used for testing are structurally separate from real family accounts and never appear in any real family's reports.
Do the curriculum numbers on this page ever change?
Yes — the curriculum keeps growing through the same write-ahead, internally-reviewed process described above. The counts on this page reflect the live curriculum as of the date noted next to them.
What happens if my child doesn't get a word right away?
Nothing gets marked wrong or held against them. A word that needs more practice just keeps quietly resurfacing in later missions and in Memory Mission reviews until it's solid — there's no fail state, timer, or streak-loss messaging anywhere in the app.
See whether the approach fits your child
Internal testing can tell us whether the software follows its rules. It cannot tell you exactly how an individual child will respond. The guided demo lets your family experience the learning approach directly before deciding whether to continue.