Skip to main content

Learning frameworks

Every AI suggestion an adult sees on the platform is grounded in a named research construct with a zone-of-proximal-development position. This page is the list. It is the same list the parent app's "Our approach" screen shows, kept here so a teacher, coach, or state reviewer can check the source.

Scarborough's Reading Rope

For decades reading was treated as one ability. It is two, braided together. Scarborough (2001) is the field-standard picture of how they grow. Reading Rockets overview.

Scarborough's Reading Rope: language comprehension strands and word recognition strands braid together into skilled reading
Two strands, braided. The Talking Book works the top strand; the Words activities cover both.
StrandWhat it isHow it is built
Language comprehensionVocabulary, background knowledge, how sentences work, reasoning about a storyMostly by talking and being read to
Word recognitionHearing the sounds inside words, matching letters to sounds, recognising whole words on sightExplicit, sequenced practice

The Words activities cover both strands; the teacher can see which one needs attention. The Talking Book works the comprehension strand.

The sound ladder

Before children read, they learn to hear smaller and smaller parts of spoken language. One rung at a time; skipping ahead is one of the strongest predictors of later reading difficulty. National Early Literacy Panel (2008), Developing Early Literacy (NICHD record).

The sound ladder: five rungs from word awareness at age two to three up to phoneme segmenting at age five
Five rungs, in order. The apps hold a child at the rung they are on.
RungSkillExampleAbout age
1Word awarenessCounts words in a sentence2 to 3
2Syllable awarenessClaps "ba-na-na"3
3Onset, rime and rhymingHears "cat / hat" as a rhyme3 to 4
4Phoneme blending/c/ /a/ /t/ becomes "cat"4 to 5
5Phoneme segmentingSays the sounds in "cat", swaps /c/ for /h/5

Overview: Reading Rockets, phonological and phonemic awareness.

Clements & Sarama learning trajectories

Children learn early maths in a known order. First "one, two, three" without meaning, then touching each object, then understanding that the last word names how many. Clements and Sarama's learning trajectories are the most-studied developmental progressions in early maths. LearningTrajectories.org. The Count activities track where the child is and nudge one rung ahead, never two.

The five counting principles

"Counting" is not one skill. It is five. Gelman & Gallistel (1978), The Child's Understanding of Number. Harvard University Press.

PrincipleIn plain words
One-to-oneOne number for each thing
Stable orderAlways 1, 2, 3, 4, 5
CardinalityThe last number is how many
AbstractionYou can count anything
Order irrelevanceStart anywhere, same total

If a child can recite numbers but does not know the last one tells you how many, that is a cardinality gap, not a counting gap, and it needs different practice. The teacher is told exactly which principle to coach.

Vygotsky's zone of proximal development

The space between "can do alone" and "too hard." Every activity the system suggests sits in it, and every suggestion is labelled so the teacher can see why it was offered. Overview.

Zone of proximal development shown as three zones: consolidate below the edge, target at the edge, stretch above it
Every activity is tagged with one of three zones for a given child.
ZoneWhere it sits
ConsolidateJust below the edge. Builds confidence and automatic recall
TargetAt the edge of what the child can do with a small nudge
StretchJust above the edge. Used sparingly, when readiness is high

What the AI does, and never does

For children from birth to five, an AI should never be positioned as a chatbot, companion, or emotional bond. UNICEF's Policy guidance on AI for children (2021) is the reference. On the platform the AI sits in the background as a coach for the teacher. The child never talks to it.

DoesNever does
Listens for hesitation and pronunciation during an activity to gauge confidenceChat with the child. No open conversation, ever
Maps each child's spot on a learning trajectoryAct as a friend, companion, or emotional support
Suggests the next activity to the teacher, never to the childGenerate freeform stories or content for the child
Sends the teacher a one-paragraph "why" with every suggestionListen when an activity is not on screen
Stays inside the school's environment. Not used to train modelsReplace the teacher's judgment

Standards the platform maps to

Where the evidence stops

These are frameworks, not outcomes. They tell us what to measure and in what order; they do not by themselves prove that the platform improves anything. Outcome evidence comes from the pilot and will be published on this site when it exists.

Sources

  1. Scarborough 2001, via Reading Rockets. https://www.readingrockets.org/topics/about-reading/articles/scarboroughs-reading-rope-groundbreaking-infographic
  2. National Early Literacy Panel 2008, NICHD. https://www.nichd.nih.gov/publications/product/346
  3. Reading Rockets, phonological and phonemic awareness. https://www.readingrockets.org/topics/phonological-and-phonemic-awareness
  4. Clements & Sarama, LearningTrajectories.org. https://www.learningtrajectories.org/
  5. Gelman & Gallistel 1978, Harvard University Press. https://www.hup.harvard.edu/books/9780674116375
  6. Vygotsky, Mind in Society 1978, ZPD overview. https://www.simplypsychology.org/zone-of-proximal-development.html
  7. UNICEF 2021. https://www.unicef.org/innocenti/reports/policy-guidance-ai-children
  8. Head Start ELOF. https://headstart.gov/interactive-head-start-early-learning-outcomes-framework-ages-birth-five
  9. NAEYC DAP 2022. https://www.naeyc.org/resources/position-statements/dap
  10. W3C WCAG 2.2. https://www.w3.org/TR/WCAG22/