Methodology
How It Works — and Why You Can Trust It
Transparent methodology. No black boxes. Here's exactly how the phoneme matrix works, how we calculate our claims, what sources we rely on, and who built it.
The Phoneme Matrix — How It Works
Every accent of English contains a unique inventory of sounds. Some of those sounds already exist in your target language — you just don't know which ones yet. The phoneme matrix is the engine that maps your accent's sounds to French, German, Spanish, and Italian.
We cross-reference every target-language phoneme against the sound inventory of eight English accents: American, British (RP), Australian, Irish, Scottish, South African, Indian, and Canadian. The result is a matrix of 560 entries containing 3,920 data points in total.
What Each Entry Contains
Every cell in the matrix records seven pieces of information for a given accent × sound combination:
1. Status — Is this sound already in your accent (known), reachable with a small adjustment (bridge), or genuinely new (new)?
2. Bridge Word — An everyday English word from your accent that contains the target sound (e.g., a Scottish speaker's "loch" for the German ach-Laut).
3. Physical Coaching — Tongue, lip, and jaw instructions tailored to your starting position.
4. Drill Sequence — A progressive set of exercises moving from isolated sound to word to sentence.
5. Common Errors — The mistakes speakers of your accent typically make with this sound, and how to avoid them.
6. Difficulty Rating — A 1–5 scale reflecting how much effort this sound typically requires for your accent profile.
7. IPA Symbol — The International Phonetic Alphabet notation for precise phonetic reference.
Sound Overlap Across Accents
The key insight is that different English accents share different amounts of overlap with each target language. A Scottish speaker already produces the rolled R that Spanish requires. An Australian speaker already makes the broad vowel that French nasals demand. A British RP speaker says "bird" with a vowel that is the German Ö.
The matrix quantifies this overlap precisely, so every learner gets a personalised path that starts from what they already know — not from zero.
How Claims Are Calculated
When we say a Scottish speaker can learn Spanish pronunciation "44% faster," that number comes from a specific, reproducible formula — not marketing intuition.
The Formula
Remaining work = (Total sounds − Known sounds) ÷ Total sounds
Example: Scottish + Spanish
Spanish has approximately 32 distinct phonemes. A Scottish English speaker already produces 18 of them natively (the rolled R, most vowels, several consonant clusters). That leaves only 14 sounds that need active learning — 44% of the total.
(32 − 18) ÷ 32 = 44% of sounds still to learn
In other words, a Scottish speaker skips 56% of the pronunciation work a generic learner faces — they only need to actively learn 44% of the sounds.
The Inputs
Sound inventory analysis — cataloguing every phoneme in each English accent using IPA transcription and phonological literature.
Difficulty weighting — not all "new" sounds are equally hard; some require a small lip adjustment, others a fundamentally new tongue position. The matrix weights accordingly.
Phonological literature — contrastive analysis research that documents which L1 sounds transfer to which L2 sounds, and where interference occurs.
Comparative calculation — the same formula applied consistently across all 8 accents × 4 languages, producing comparable percentages.
Limitations
We believe in transparency about what these numbers are — and what they are not.
- •These are estimates, not clinical measurements. The percentages indicate directional advantage based on sound overlap — they are not the output of a randomised controlled trial.
- •Individual variation matters. Your musical ear, motivation, practice frequency, and prior language exposure all affect real-world learning speed.
- •We welcome academic validation. If you're a phonetician or applied linguist interested in testing these mappings, we'd love to hear from you — get in touch.
Sources and Foundations
The phoneme matrix is not generated by AI. It is hand-researched, drawing on established phonological scholarship:
International Phonetic Alphabet (IPA)
The global standard for transcribing speech sounds, maintained by the International Phonetic Association. Every sound in the matrix is catalogued using IPA notation.
J.C. Wells — Accents of English (3 volumes, 1982)
The foundational reference work on English accent variation worldwide. Wells' detailed phonetic descriptions of regional accents form the basis of our accent inventory analysis.
Contrastive Analysis Literature
The branch of applied linguistics that systematically compares the phonological systems of two languages to predict areas of difficulty and transfer. This research informs which sounds map, which bridge, and which require new learning.
L2 Phonological Acquisition Research
Studies on how adults acquire new sound systems, including Flege's Speech Learning Model and Best's Perceptual Assimilation Model, which describe how learners perceive and produce non-native sounds through the filter of their L1.
A Note on AI's Role
AI powers the conversational personality of your tutor, the session pacing, and adaptive review scheduling. It does not generate the phonetic content. Every sound mapping, bridge word, coaching cue, and difficulty rating is derived from phonological research — not from a language model.
About the Founder
My Accént was created by an Australian-origin attorney who has practised law in Switzerland for over 25 years. Living and working across English, German, French, and Swiss-German gave him a front-row seat to how accents shape language learning — and how rarely anyone accounts for it.
The idea crystallised watching his daughter Ylva grow up in Zürich. By the time she started school, she was navigating English, High German, French, and Züridütsch — the local Swiss-German dialect that became her fifth language. He noticed she wasn't "learning" new sounds so much as recognising ones she already had. The sounds in her mouth were the bridge.
That observation — that your existing sounds are the fastest route into a new language — became the foundation of My Accént. Züridütsch remains the unofficial fifth language in the app, a nod to where the idea was born and a reminder that every accent, no matter how local, carries hidden advantages.
Frequently Asked Questions
How does the accent advantage matrix work?▾
Are the pronunciation time estimates accurate?▾
What sources underpin the phonetic data?▾
Does AI generate the pronunciation advice?▾
Who created My Accént?▾
See the Matrix in Action
Explore the full accent × language matrix yourself — or take the accent quiz to discover your personal sound advantage.