Some learners shadow daily and sound increasingly like the recording, until a new topic leaves them silent. Some speak freely every day but preserve the same errors. Others collect explanations and spend almost no time speaking.
The problem may not be the method. It may be that the whole week contains only one kind of practice. Paul Nation's Four Strands offers a useful audit: meaning-focused input, meaning-focused output, language-focused learning and fluency development.
What each strand contributes
Meaning-focused input: Listen or read with most of your attention on the message and a high level of comprehension. For example, understand why two people in a short dialogue change their plans.
Meaning-focused output: Speak or write to communicate something that stretches your current resources. Explain the same change of plan to a friend without a full script.
Language-focused learning: Deliberately study a chunk, grammar feature or pronunciation contrast. Learn *Would it work if we moved it to Friday?* and pay attention to its reductions and pauses.
Fluency development: Use familiar language and familiar content with some pressure to perform faster or more smoothly. Explain the change of plans in 90, 60 and 45 seconds.
Nation recommends roughly equal time for the four strands across a balanced course. This is a course-design framework, not an experimental law requiring exactly 25% in every study session. Audit your week before reaching for a stopwatch.
Four 12-minute sessions around one situation
Use "weekend plans have changed" throughout the week so each strand supports the others.
Session one: understand a message
Choose an accessible dialogue. Listen twice and answer: who changed the plan, why, and what did they decide? Give a three-sentence summary. Look up the few essential unknown words, but do not turn every line into dictation.
Session two: learn language you can carry away
Select two useful expressions, such as *Something has come up.* and *Could we do it another day?* Hear and shadow them, then replace the reason and time with your own. Deliberate study should return to communicative use, not stop at terminology.
Session three: create a complete output
Keep only three prompts: reason, proposed time and checking the other person's availability. Record for 60–90 seconds. Select one communication problem, compare a model and answer again.
Session four: become fluent with familiar content
Repeat the same message in 90/60/45-second rounds. Do not introduce new vocabulary or ideas. Reduce long searches, preserve the key information and connect the sequence more clearly.
A strand depends on what you are actually doing
One recording can serve different purposes. A speech you barely understand is not ideal meaning-focused input because the comprehension load is too high. Analysing its grammar line by line belongs to language-focused learning instead.
Reading an unfamiliar text increasingly fast is not necessarily fluency practice either; the language and content are not yet familiar.
Fluency work needs four broad conditions: largely familiar language and content, attention on communicating meaning, some pressure or encouragement to go faster, and enough input or output. Speed alone is not the definition.
Three common imbalances and their fixes
Only input: Add a small product to each resource. After listening, explain the message to an imagined colleague for 60 seconds.
Only output: Choose two or three recurring problems each week for focused study, then return to the original task with the model you learned.
Only language study: Add complete messages and timed retelling of familiar content. Being able to analyse a sentence is not the same as being able to use it in sequence.
With only fluency work, you may deliver old material smoothly without adding new language resources. Input and focused study keep those resources growing.
Where MirrorSay belongs
MirrorSay is most useful in language-focused learning: say your own sentence, compare a natural formulation, hear the model and compare timing. Then return the repaired sentence to meaning-focused output or fluency practice, so one successful shadowed line does not stand in for the whole speaking curriculum.
For the timed procedure, see The 4/3/2 Speaking Drill. To organise attempts, targeted input and evaluation around a real outcome, read Production-Oriented Approach for Solo Speaking.
References
- Nation, P. (2007). The Four Strands. *Innovation in Language Learning and Teaching*, 1(1), 2–13. DOI · Author-hosted PDF
*MirrorSay is a free AI speaking-practice app: say a sentence, see a natural way to express it, then shadow the model. No sign-up, and speech is processed on-device whenever possible.*