The Shadowing Technique: A Practical Guide for Language Learners

Shadowing feels awkward for the first few minutes. Then a passage you could only understand becomes one you can produce.

What Shadowing Is

Shadowing is saying what a speaker says while they are still saying it, copying the rhythm and melody. The defining feature is the overlap: you start on the first syllable and trail the speaker by under a second, tracking sound in real time instead of reproducing a sentence from memory. Not to be confused with:

• Listen-and-repeat: the speaker finishes, then you speak. Works from memory, a different task.
• Parallel reading (chorusing, text-present shadowing): shadowing while reading the script. The stage to start with.
• Reading aloud: text, no audio model, nothing corrects your rhythm.
• Mumbling: shadowing under your breath. A warm-up.

You cannot shadow a word you have not heard, so shadowing exposes at once which sounds and which stretches of connected speech you are guessing at.

Where It Comes From

In 1973 the psycholinguist William Marslen-Wilson described “close shadowers” in Nature: people who repeat connected prose 250 to 300 milliseconds behind, about one syllable. They still corrected errors in the text, so they were processing meaning, not echoing sounds. Interpreter schools adopted the task for beginners. Sylvie Lambert’s 1992 definition still stands: a paced, auditory tracking task that involves the immediate vocalization of auditorily presented stimuli.

Ken’ichi Tamai brought it into Japanese English classes in the 1990s; Shuhei Kadota and Yo Hamada built most of the evidence below. Alexander Arguelles popularized the self-study version (walking briskly, shadowing an Assimil lesson by ear, then with translation and text, then with the text alone). He did not invent it, whatever blog posts claim.

What the Research Shows

• Listening. Hamada (2016, 43 students): better phoneme perception at both levels tested, better comprehension only in the lower-level group. A 2025 systematic review agrees: listening gains concentrate in beginners and lower-intermediates.
• Pronunciation and fluency. Whitworth and Rose (2025), 44 studies: all eight that measured fluency found gains; comprehensibility and prosody mostly improved; individual sounds inconclusive. Foote and McDonough (2017): eight weeks, better comprehensibility and fluency, accent unchanged, no control group. Shao, Saito and Tierney (2023): twelve half-hour sessions with a control group, gains in comprehensibility and accentedness, larger for learners with better auditory processing.
• How many passes. Shiki and colleagues (2010): reproduction levels off after four or five tries. Hamada: five or six passes, then move on.
• How fast. Sumiyoshi (2019): audio sped up over eight weeks improved comprehension and dictation at slow and fast speeds. Do not stay slow forever.
• Text or no text. Hamada (2021, 96 students): three rounds on an unfamiliar accent without a script did not help; reading the script between rounds two and three did.
• Understanding first. Hamada (2014): only the group that studied the passage’s content before shadowing improved.
• Not shown. No meta-analysis as of 2026. Studies are small (16 to 96 participants), a few weeks long, mostly Japanese university English classes, rarely with a delayed test. “Boosts working memory” rests on one four-week study with mixed results. “Sound native” and “walking oxygenates the brain” have no evidence.

The Practice Sequence

One passage of 30 seconds to 2 minutes of audio, five or six passes:

1. Understand it first. Read with the translation until every sentence makes sense. The step most people skip, and the one that matters most.
2. Listen once without speaking. Notice pauses, speed, rises and falls.
3. Mumble along once or twice, rhythm only.
4. Shadow with the text, a beat behind. Slow the audio if you cannot keep up.
5. Shadow without the text. Hide the translation, then the original. Two or three passes. This is where listening improves.
6. Shadow for meaning. One pass at normal speed, attending to what the passage says (content shadowing).

10 to 15 minutes, two or three times a week. Change the passage when it feels easy, raise the speed when the slow one does. Record yourself sometimes: Hamada (2015) found self-monitoring beat a partner’s feedback. Play the original and your recording back to back, fix one thing.

Common Mistakes

• Waiting for the sentence to end. That is listen-and-repeat. Start on the first syllable.
• Shadowing what you do not understand. Fix the meaning first.
• Passages over two minutes. You tire before the fifth pass.
• Whispering. Speak at conversational volume.
• Staying at half speed. Slow audio is a ramp, not a home.
• Never removing the text. The ear-only passes are the exercise.

Why a Book Is Good Material

Podcasts and video fail the first rule for everyone below advanced: you rarely understand them fully, and you cannot replay one sentence. A book with a parallel translation and sentence-by-sentence audio fixes both: understanding takes a minute, the narration is clear and even, and you replay one sentence as often as you like at the speed you choose. Narration also carries the full stress and intonation of the language without the half-finished sentences of conversation.

Still building your first few thousand words? Start with short texts by level, then move to books when the A2 and B1 stories feel easy.

Shadowing in Lingo7

Lingo7 reads books with a parallel translation in 49 languages, with sentence-level audio in 46.

1. Read the page first. Tap a sentence for its translation, a word for its card.
2. Turn on Audio mode from the top-right menu. A play button appears before every sentence, plus a floating player.
3. Shadow sentence by sentence. Tap a sentence’s play button: it plays that sentence and stops, each word lighting up as it is spoken.
4. Loop a hard sentence. Press and hold its button until the ring fills: it plays several times and counts down. Three by default; Repeats in Audio settings sets 1, 3 or 5. Tap to stop.
5. Echo pause (Audio settings) makes the voice wait after each sentence, as long as the sentence or one and a half times longer. Strictly listen-and-repeat, useful for first rounds on a long sentence. Turn it off once you keep up.
6. Slow the voice with the speed control in the floating player, 0.6 to 1.4, then bring it back up.
7. Hide the translation and shadow the same sentences by ear.
8. Shadow the whole page. Press play in the floating player: narration runs sentence after sentence, highlights the current word and turns the page. Look away and keep up.

Every catalog book in those languages has audio. Start with a page you have already read.

The Bottom Line

Shadowing is a fluency and rhythm drill with decent evidence: strongest for beginners’ listening and for everyone’s fluency and intonation. Only on material you understand, in short passages, five or six passes, from text to ear and from slow to normal speed. Ten minutes a few times a week with a book you have already read.

It pairs with reading while listening, the same sound-to-meaning link from the receptive side, and it is the third technique in our plan for learners who can read but cannot follow speech.