1. A myth that refuses to die
In my workshops over the years, one myth just refuses to die… the idea that teaching the EPI way means using Sentence Builders. Some individuals even call it ‘The sentence builders method’!
Let me tell you — it’s simply not true.
Sentence Builders are a tool. A very powerful one, sure — maybe even revolutionary in the hands of a reflective teacher — but still just that: a tool!
EPI (Extensive Procesing Instruction) isn’t about grids or colours or boxes. It’s about how learners process language… how they encode meaning, notice form, consolidate memory traces, and build fluency through repeated, meaningful, deliberate rehearsal of language.
You could teach the EPI way perfectly well using listening sequences, classroom interactions, cue cards, or even a few well-designed slides — if you understand and apply its underlying principles.
In this post I intend to debunk this myth and to propose viable alternatives for those EPI teachers who may want to start the modelling phase in different ways. These alternative techniques do not necessarily rule out the use of sentence builders if one is still keeon on using them, of course.
2. EPI as Process-Based Instruction
At its heart, EPI is a process-based instructional model — and that’s not a slogan, it’s a paradigm shift. It means focusing less on what learners produce and more on how they process input on the way there.
Over the past ten years of EPI advocacy, I’ve seen that the best teachers are those who work deliberately on the micro-skills of each language domain:
- Listening: discriminating sounds, chunking meaning units, predicting what comes next, reconstructing messages.
- Reading: decoding, inferring, processing chunks, integrating information across sentences.
- Speaking: retrieving and adapting chunks rapidly, automatizing pronunciation and syntax patterns, monitoring accuracy of output online.
- Writing: planning, retrieving, sequencing chunks, reformulating, expanding structures with increasing complexity, monitoring accuracy.
In my observation, great teachers don’t leave those micro-skills to chance. They design activities that deliberately recyle and rewire those mental proceses. Every listening, reading, speaking, or writing task becomes a laboratory of cognitive rehearsal. As my co-author, Dylan Vinales, once said: “The EPI classroom is, in essence, an ecosystem of automatisation — not a collection of worksheets”.
3. The MARS EARS framework — with or without Sentence Builders
The MARS EARS sequence is where EPI becomes visible. It takes learners from exposure to spontaneity through a very deliberate chain:
- Modelling – Rich, meaning-bearing input (spoken, written, visual).
- Awareness – Learners notice and connect form with meaning.
- Receptive Processing – Intensive listening or reading tasks demanding discrimination and recall.
- Structured Production – Controlled oral and written manipulation of chunks. In this phase, Sentence Builders should be used only to ‘warm up’ the students before retrieval practice tasks, to support the weaker ones or to check whether the answers are correct.
- Expansion – Explicit grammar teaching, Recombination and Extension of what’s been learned.
- Autonomy/Assessment – Low-stake assessment, diagnosis and consolidation.
- Review & Spontaneity – Recontextualised recycling leading to automatizion.
Nowhere in that process does it say “use a Sentence Builder.”
You can model, prompt, and scaffold language through a dozen other means — as long as the sequence keeps its cognitive gradient from receptive to productive.
4. The power of chunks and structural priming
EPI draws heavily on usage-based linguistics — the idea that language emerges through repeated procesing of meaningful lexicogrammatical chunks.
Chunks are the brain’s way of simplifying life. They allow fluent speech and comprehension by bundling grammar and lexis together into reusable units.
When learners repeatedly encounter patterns like “I’m going to…”, “Can I have…?”, or “On the weekend I usually…”, something powerful happens: structural priming.
As Bock (1986) and Pickering & Branigan (1998) showed, exposure to a structure increases the likelihood that we’ll use it. That’s why, in my experience, students who are bombarded with the same high-frequency patterns in different contexts start producing them effortlessly. It’s not magic — it’s cognitive economy at work.
But in EPI grammar is also taught explicitly, through deductive teaching or inductive learning, as the teacher feels fit, in the Expansion phase of the MARSEARS sequence. No sentence builders is necessary in this phase either, unless you want to make it easier for the students, that is…
5. Listening as modelling: the interpersonal gateway
One of the things I’m proudest of in my work on EPI is the idea of interpersonal listening-as-modelling.
Traditional listening exercises test comprehension — “Did you catch the right answer?” — and that’s it.
EPI flips that.
In my approach, learners don’t just listen; they use what they hear as a model for their own output.
They mirror, adapt, and reuse the same structures almost instantly.
For example, after a short teacher–student exchange about weekend routines, I might ask students to replicate it with a small twist — change the activity, the day, or the person.
This transforms listening from a passive test into a productive rehearsal — the most efficient form of input processing I’ve seen in decades of teaching.
6. The engine of learning: repeated and varied processing
Every teacher knows that repetition matters… but not all repetition is equal.
Research from Craik & Lockhart (1972) to Webb (2021) confirms that depth and variety of processing are what truly consolidate learning.
In EPI, repetition is not parroting. It’s a cycle of re-encoding: hearing, noticing, matching, reformulating, expanding.
Each re-encounter engages a different network in the brain, reinforcing long-term memory traces and reducing cognitive strain.
Unfortunately, many classrooms stop too soon. They introduce a structure once, practise it briefly, then move on.
But language doesn’t work like that.
In EPI, we recyle chunks over multiple modalities and days — always slightly recontextualised, always processed anew.
7. How You Can Present and Model the Target Chunks Without Sentence Builders
Modelling is where it all begins. And no, you don’t need a grid. You just need clarity, intention, and a bit of creativity.
Here are some of my favourite ways to model chunks without ever touching a Sentence Builder:
1. Flashcards for rapid retrieval and noticing
Create flashcards with the target chunks on one side and visuals or translations on the other.
Start with teacher-led modelling: flash a card, say the phrase aloud, and have learners repeat with rhythm and gesture.
Then move to pair quizzing: one learner shows the card, the other produces the chunk aloud, swaps roles, and times each other.
Later, combine several flashcards into quick oral chains (“On Saturday I go swimming + with my friends + in the morning”), encouraging fluency through speed and combination.
Flashcards keep retrieval active, reduce cognitive load, and strengthen form–meaning mapping through constant low-stakes practice.
All these approaches maintain what EPI values most: clarity of modelling, deliberate recycling, and deep engagement with meaning… and not a single Sentence Builder in sight.
2. Listening-based visual anchoring
Display simple visuals — icons, pictures, emojis — while reading or playing short dialogues aloud.
Ask learners to match what they hear, spot faulty descriptions, or order the visuals.
The picture set becomes your “invisible Sentence Builder.”
3. Guided micro-dictations
Deliver short, chunk-rich dictations like “On Saturdays I go to the gym.”
Learners reconstruct, translate, and underline patterns like “time phrase + verb + place.”
Simple, effective, memorable.
4. Cued oral modelling with substitution
Use cue cards with categories (time, activity, place).
Model examples while changing one cue at a time:
“On Saturday I play football.”
“On Saturday I play tennis.”
“On Sunday I go swimming.”
Learners repeat and adapt — automatisation through meaningful variation.
5. Listening-as-modelling chains
Play a short recording or give a live model. Learners shadow, then reproduce it with one altered detail.
This keeps listening and speaking intertwined — a direct substitute for the Sentence Builder grid.
6. Faulty-translation and reformulation games
Project slightly wrong English translations of accurate TL sentences.
Learners spot and fix them — deep procesing guaranteed.
7. Collaborative classification
Give learners printed chunk strips.
Ask them to group by meaning (activities, times) or syntax (verb + infinitive / verb + noun).
The tactile act of sorting helps internalise structure — cognitive learning by touch and sight.
All these approaches maintain what EPI values most: clarity of modelling, deliberate recycling, and deep engagement with meaning… and not a single Sentence Builder in sight.
8. Interactive oral modelling
Start with a natural class exchange.
“Do you go out on weekends?”
“Yes, I usually go jogging.”
“You usually go jogging? Great! Who with?”
“With my friends.”
Each question–answer pair repeats the target chunks in meaningful variation. The teacher subtly primes the syntax while keeping communication real.
Please note: none of the above excludes the use of Sentence Builders as chunk organisers, scaffolding tools, or as a means to reinforce and consolidate the initial modelling achieved through the techniques described above.
8. EPI as a mindset — and why Sentence Builders still matter
Now, let me be very clear. EPI is a mindset, not a method locked into a specific tool.
It’s about cognitive design, not graphic design.
What defines an EPI teacher, in my experience, is not the slide they show but the mental operations they induce in their learners.
Sentence Builders can do that — but so can good listening chains, reformulation tasks, and carefully scaffolded dialogues.
That said, there is a very solid theoretical reason I’ve advocated Sentence Builders for years.
They work because they align beautifully with John Sweller’s Cognitive Load Theory, particularly his concept of worked examples. Sweller (1988, 1994) showed that novices learn faster and more efficiently when given clear, structured models to study — reducing extraneous cognitive load.
Sentence Builders are, in essence, linguistic worked examples. They make syntax and lexis visible, freeing up working memory so learners can focus on meaningful processing rather than rule-hunting or guessing.
As I often say, they are not crutches — they are cognitive prosthetics, temporary but invaluable.
Once the learner’s brain has internalised the patterns, the scaffold can — and should — fade.
In ten years of EPI advocacy, I’ve seen this countless times: teachers who overuse Sentence Builders create dependency… those who understand their cognitive purpose create fluency.
Concluding remarks
So, no — you don’t need Sentence Builders to teach the EPI way.
What you do need is a deep understanding of:
- Process-Based Instruction: deliberately targeting micro-skills and scaffolding processing across all four skills.
- Chunk-Based Learning: building fluency through patterned input and structural priming.
- Interpersonal Listening-as-Modelling: making listening the springboard for output.
- Repeated, Varied Processing: recyling language across contexts until it becomes automatic.
- MARS EARS: a flexible yet principled roadmap for cognitive sequencing.
If these are in place, you are already teaching the EPI way — Sentence Builder or not.
Because in the end, the true magic of EPI doesn’t lie in the layout of a grid…
but in the architecture of the mind it helps construct.
The Language Gym 