{"slug":"sound-to-phonogram","title":"Sound-to-phonogram","date":"2026-03-10T10:13:59.845Z","excerpt":"The reverse of phonogram-to-sound mapping","tags":["project"],"order":53,"published":true,"content":"\n\n\n\n\n\n\n\n

You were taught to read. Nobody taught you to spell — not really.

Sure, there were lists to memorise and tests to pass. But if you've ever frozen mid-sentence trying to remember whether it's recieve or receive, you know exactly how shallow that learning was. The words got into your head, but the system behind them never did.

That's not a personal failure. It's a design flaw in how literacy has traditionally been taught.

For decades, the default approach has been to show learners a word and ask them to decode it — letter by letter, left to right. It works well enough for reading. But reading is just recognition. It doesn't force the brain to do the hard work of truly owning a word. Spelling does. And spelling has largely been an afterthought.

Sound-to-Phonogram is an attempt to fix that. It flips the sequence — starting with sound instead of print — and in doing so, it turns out to teach both skills better than the traditional method teaches either. The result isn't just better spelling. It's a deeper, more durable understanding of how the English language actually works.

\n\n\n\n\n

This document explains why, and how.

Part 1: Pedagogical Foundations

1.1 Why Spelling Builds Literacy Better Than Reading

To build a high-performance literacy tool, we must move past the \"common sense\" idea that reading and spelling are just two sides of the same coin. They are neurologically distinct processes that require different levels of cognitive effort.

Encoding (Speech-to-Print) — The Educational Exercise

Encoding is the act of constructing a visual representation of a sound. When a learner is asked to spell a word, their brain is forced into a high-friction, multi-step mechanical process:

Phonological Isolation: Hearing the whole word and segmenting it into individual phonemes (sounds).
Working Memory Loading: Holding those abstract sounds in order while searching for their visual counterparts.
Orthographic Retrieval: Recalling the specific grapheme (letter pattern) rules that represent those sounds in a particular context (e.g., knowing /eɪ/ in the middle of a word is usually ai but at the end is ay).
Sequential Execution: Mapping those symbols in the correct visual order.

Key Insight: This \"Cognitive Friction\" is exactly what forces Orthographic Mapping. The brain bonds the spelling, pronunciation, and meaning of the word into long-term memory — this is the Construction of Knowledge.

Decoding (Print-to-Speech) — The Assessment Tool

Decoding is the act of recognizing a visual string and retrieving its sound.

Rapid Retrieval: If a word is already mapped, decoding happens in milliseconds via the \"lexical route.\" It feels effortless.
The Guessing Trap: If a word has not been mapped, the brain often bypasses phonetic logic and resorts to guessing based on first letters, word length, or context.

Key Insight: Reading is a Retrieval Task. It does not build the neural pathway; it merely exercises one that should already exist. Because it allows for guessing, it is an unreliable way to teach spelling — but as a teacher, it provides a perfect window into what the student has actually internalized.

1.2 The Speech-to-Print Methodology

Because of the asymmetry described above, the system completely abandons the traditional \"reading flashcard\" model in favor of a Speech-to-Print methodology.

Approach A: Print-to-Speech (Traditional / Assessment Route)

Mechanism: Look at letters → Sound them out.
The Flaw: In an opaque language like English (where one grapheme like ough has many sounds), this forces learners to memorize an overwhelming number of \"exceptions.\" It prioritizes Passive Recognition.

Approach B: Speech-to-Print (Encoding / Educational Route)

Mechanism: Hear the sound → Map the letters.
Rationale: The human brain is biologically hardwired for speech (evolutionary), but not for reading (cultural). By starting with the sound — which the learner already understands — and showing the statistical probability of how that sound is spelled, the language becomes a Logical Code.

The Pedagogical Golden Rule: If you can spell a word, you can almost always read it fluently. The reverse is not true. By focusing 100% on encoding, we guarantee the student achieves both literacy goals simultaneously.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

Feature	Print-to-Speech (Decoding)	Speech-to-Print (Encoding)
Primary Function	Assessment of existing knowledge	Active Exercise to build knowledge
Cognitive Load	Low (allows for guessing/shortcuts)	High (forces orthographic mapping)
Starting Point	Artificial visual symbols (letters)	Natural spoken language (sounds)
Impact on Memory	Tests rapid retrieval/fluency	Constructs permanent mental links

1.3 The Surgical Transcriber: Diagnostic & Stress-Free Assessment

The Surgical Transcriber (luyendoc.io.vn/transcriber) is integrated not as a reading aid, but as a diagnostic and stress-free assessment tool.

Lowering the Affective Filter

Traditional testing creates \"Test Anxiety,\" which hijacks working memory and causes students to underperform. The Transcriber enables Formative Assessment:

The student reads a real-world text (from the 1,200+ article library).
Because IPA is only placed over counter-intuitive graphemes, the student feels supported rather than tested.
Teacher insight: If a student reads the un-annotated parts fluently, you have proof of mastery without ever needing a high-pressure exam.

Teacher-Annotated Diagnostics

Rationale: When a student mispronounces a word, the teacher can annotate the error directly — differentiating between a phonological error (can't hear/produce the sound) and an orthographic error (doesn't know the spelling rule).
Closing the Loop: These annotations direct students back to a specific route in the Spiral Curriculum. A missed ch as /k/ points back to the \"Consonant Traps\" module.

The Unified Learning Loop

The Gym (Encoding Curriculum): The student builds the orthographic map via Speech-to-Print exercises.
The World (Surgical Transcriber): The student applies that knowledge in a stress-free reading environment.
The Coach (Teacher Annotations): The teacher uses the Transcriber as a diagnostic window and sends students back to the Gym for targeted training.

Part 2: Adaptive UI & Technical Architecture

2.1 The Five-Level Difficulty Progression

Instead of relying on a backend algorithm to track user history over time, the application uses a dynamic, stateless UI progression. The interface shifts between cognitive difficulty levels based on the Route Configuration (set by the teacher's curriculum) and the Session Escalator (driven by the user's immediate success streak).

Level 0–2 — The Scaffold (Multiple Choice): Displays the word with the target grapheme hidden. Presents 3 options: 1 correct grapheme and 2 phonetically related distractors. Focuses on initial pattern recognition without overwhelming working memory.
Level 3 — The Bridge (Constrained Input): Removes multiple-choice buttons. Displays a text input with visual character constraints matching the exact letter count of the missing grapheme (e.g., [ _ _ _ ] for igh). Forces internal retrieval while preventing blank-page paralysis.
Level 4 — The Test (Unconstrained Input): Displays a single, uniform blank space for the missing grapheme, regardless of length. Requires pure internal retrieval — proving the learner understands phonogram representation in morphological context.
Level 5 — Mastery (Pure Dictation): Presents only audio pronunciation and a semantic anchor. Requires the user to type the entire word from scratch. Proves complete phoneme-to-grapheme segmentation.

The Session Escalator (Stateless Adaptation): If a user at Level 0 answers 3 questions correctly in a row, the Svelte component automatically upgrades the next question to Level 3. A single mistake instantly degrades the UI back to Level 0 to repair the pathway.

2.2 UI/UX Design Principles

Phonetic Anchoring: Every interaction begins with high-quality audio pronunciation paired with the IPA transcription to ground the abstract task in the primary biological system (speech).
Layout Optimisation: The 3-option multiple-choice phase is strictly maintained to prevent cognitive overload and ensure clean, mobile-friendly touch targets without vertical scrolling.
Granular Feedback Loops: Error states do not simply mark a word \"wrong.\" The UI explicitly breaks down the phonetic error by striking through the incorrect pattern and highlighting the correct phonogram block, ensuring teaching occurs at the exact point of failure.

2.3 Technical Stack

Frontend Framework: Svelte components utilising reactive stores and conditional rendering to seamlessly swap UI interfaces (Multiple Choice ↔ Input ↔ Dictation).
Application Logic: TypeScript state machine managing penalty/promotion thresholds, dynamic distractor generation based on shared phonemes, and immediate feedback loops entirely on the client side.
Data Structure: A deeply categorised, static JSON array of routes. No backend user database is required — making the application instantly accessible, highly scalable, and completely privacy-compliant.

Part 3: The 40-Route Spiral Curriculum

3.1 Curriculum Design Philosophy

Rather than relying on a backend algorithm or a rigid phoneme count, the system uses a Teacher-Curated Spiral Sequence driven by orthographic friction rather than phonological symmetry.

Blocked-to-Interleaved Pipeline: Concepts are introduced via Blocked Practice (lessons focusing on one sound) to build initial pattern recognition, then immediately tested via Interleaved Practice (reviews mixing multiple sounds) to force long-term retention.
Friction-Driven Mapping: Because 84% of English consonants map reliably 1:1, they are excluded to save curriculum space. The sequence expands dynamically to address high-friction vowel ambiguities.

3.2 Module 1 — Long Vowels (Complex Teams)

Route 1: Long E /iː/ — [Default UI: Level 0–2]
Route 2: Long A /eɪ/ — [Default UI: Level 0–2]
Route 3 (Micro-Review): Long E vs. Long A — [Default UI: Level 3]
Route 4: Long I /aɪ/ — [Default UI: Level 0–2]
Route 5: Long O /oʊ/ — [Default UI: Level 0–2]
Route 6 (Micro-Review): Long I vs. Long O — [Default UI: Level 3]
Route 7 (Cumulative Review): Long E, A, I, O — [Default UI: Level 3]
Route 8: Long OO /uː/ & Long U/Yoo /juː/ — [Default UI: Level 0–2]
Route 9 (Micro-Review): Long OO/U vs. Long O — [Default UI: Level 3]
Route 10 (Mega-Review 1): All Part I Long Vowels — [Default UI: Level 4]

3.3 Module 2 — Short Vowels (Core Anchors)

Route 11: Short I /ɪ/ — [Default UI: Level 0–2]
Route 12: Short E /ɛ/ — [Default UI: Level 0–2]
Route 13 (Micro-Review): Short I vs. Short E — [Default UI: Level 3]
Route 14: Short A /æ/ — [Default UI: Level 0–2]
Route 15 (Cumulative Review): Short I, E, A — [Default UI: Level 3]
Route 16: Short U /ʌ/ — [Default UI: Level 0–2]
Route 17: Short OO /ʊ/ — [Default UI: Level 0–2]
Route 18 (Micro-Review): Short U vs. Short OO — [Default UI: Level 3]
Route 19 (Mega-Review 2): All Part II Short Vowels — [Default UI: Level 4]
Route 20 (Master Review A): Interleaved Long vs. Short Vowels — [Default UI: Level 4]

3.4 Module 3 — Other Vowels, Diphthongs & R-Controlled

Route 21: Broad A /ɑː/, Short O /ɒ/, AW /ɔː/ — [Default UI: Level 0–2]
Route 22: The Diphthongs /aʊ/ & /ɔɪ/ — [Default UI: Level 0–2]
Route 23 (Micro-Review): Broad Vowels vs. Diphthongs — [Default UI: Level 3]
Route 24: Stressed R /ɜːr/ — [Default UI: Level 0–2]
Route 25: R-Controlled /ɔːr/ — [Default UI: Level 0–2]
Route 26 (Micro-Review): /ɜːr/ vs. /ɔːr/ — [Default UI: Level 3]
Route 27 (Cumulative Review): Broad, Diphthongs, Stressed R — [Default UI: Level 3]
Route 28: R-Controlled Arrays (/ɛər/, /ɪər/, /aɪər/, /jʊər/) — [Default UI: Level 0–2]
Route 29 (Micro-Review): All R-Controlled Vowels — [Default UI: Level 3]
Route 30: The Schwa /ə/ — [Default UI: Level 0–2]
Route 31 (Micro-Review): Schwa vs. Short Vowels — [Default UI: Level 3]
Route 32 (Mega-Review 3): All Part III Vowels — [Default UI: Level 4]

3.5 Module 4 — The Consonant Traps

Route 33: The /k/ Trap (c, k, ck, ch, que) — [Default UI: Level 0–2]
Route 34: The Palatal Trap /ʃ/ & /ʒ/ — [Default UI: Level 0–2]
Route 35 (Micro-Review): The /k/ Trap vs. Palatals — [Default UI: Level 3]
Route 36: The Affricate Trap /tʃ/ & /dʒ/ — [Default UI: Level 0–2]
Route 37: The \"F & S\" Illusions /f/ & /s/ — [Default UI: Level 0–2]
Route 38 (Mega-Review 4): All Consonant Traps — [Default UI: Level 4]

3.6 Module 5 — The Global Crucible (Ascension to Mastery)

Route 39 (Crucible 1): Global Vowel Randomisation — [Default UI: Level 4]
Route 40 (The Summit): Pure Dictation across the full database — [Default UI: Level 5]

\n\n"}