Shadowing: the language learning technique explained

Shadowing is speaking along with a recording, a fraction of a second behind the speaker, without pausing the audio. You do not wait for the sentence to end. You chase it. In Chinese the technique is called 跟讀 (跟读) gēndú, "follow-reading", and in Japanese シャドーイング shadōingu.

It is a pronunciation and listening drill, and a good one. It is also frequently sold as a complete method, which it is not. This page explains how to do it, what it trains, where it stops working, and how it differs from its two quieter relatives, chorusing and echoing.

What shadowing is and where it comes from

Shadowing began as a training exercise for simultaneous interpreters, who have to listen and speak at the same time for a living. In the 1990s Japanese researchers, starting with Ken Tamai, brought it into English classrooms as a listening exercise, and it has been a staple of language teaching in Japan ever since.

Among self-learners the technique is tied to one name: Alexander Arguelles, who turned it into a daily routine and demonstrated it on video. Most of what the internet says about shadowing is a diluted version of his instructions.

The defining feature is the overlap. Your voice and the recording run together, with a lag of a word or two. That is what separates shadowing from ordinary listen-and-repeat.

How to do shadowing step by step

1. Pick a short clip you understand. Thirty seconds to two minutes, one speaker, natural speed, with a transcript. Textbook dialogues, podcast segments for learners and audiobook passages all work. If you cannot follow the meaning, you are practising noise.
2. Listen first. Play it two or three times with the transcript and look up anything unclear. This is ordinary comprehensible input work, and it comes before the drill.
3. Mumble along. Play the clip and follow under your breath, eyes on the text. Aim for rhythm and pitch, not perfect sounds.
4. Shadow out loud with the text. Full voice, a beat behind the speaker. Use one earbud or keep the volume moderate so that you can hear yourself.
5. Shadow without the text. This is the real thing. You will drop syllables. Keep going and catch the next phrase.
6. Record one pass and compare. Listen to yourself against the original and pick one thing to fix: a tone, a vowel, a place where you rushed.
7. Repeat the same clip over several days, then replace it.

Ten to fifteen minutes is plenty. Shadowing demands full attention, and after a quarter of an hour most people are no longer imitating, only making noise in the general direction of the language.

What shadowing trains and what it does not

Shadowing trains the things that are hard to learn from a description: rhythm, intonation, speed, linking between words, and for tonal languages the way tones behave in a real sentence. It also sharpens listening, because you cannot repeat what you did not catch.

The research broadly agrees. A 2024 systematic review, written as a master's dissertation at the University of Oxford, looked at 44 studies and found that shadowing can make learners easier to understand and improves fluency and prosody. For individual sounds the evidence was inconclusive, and the review notes that many of the studies had weak designs.

That matches the limits you will notice yourself:

• It does not fix sounds you cannot hear. If you perceive Mandarin x and sh as the same sound, you will shadow them as the same sound, fluently. Individual sounds and tones need separate ear training and feedback first. See pronunciation and tones.
• It does not teach you to build sentences. You are borrowing someone else's words. Conversation still needs conversation.
• It can cement errors. Nobody corrects you while you shadow. Recording yourself, or showing a recording to a tutor now and then, is what stops a mistake from becoming a habit, the slow hardening described under fossilization.
• It is poor vocabulary study. New material arrives too fast to remember.

Chorusing and echoing: how they differ from shadowing

Two related techniques solve the main weakness of shadowing, which is that everything goes by once and too quickly. Each has its own page, chorusing and echoing, and both sit with the other methods on this site in the methods directory.

	Shadowing	Chorusing	Echoing
What you do	Speak along with running audio, slightly behind	Loop one short phrase and speak exactly in time with it, again and again	Listen to a short phrase, pause, replay it in your head, then say it
Length of audio	30 seconds to a few minutes	One phrase, 3 to 15 seconds	One phrase of a few words
Timing	A beat behind the speaker	Simultaneous	After a silence
Repetitions	A few passes	Dozens per phrase	Several per phrase
Best for	Rhythm, flow, listening stamina	Precise prosody and muscle memory	Hearing accurately before you speak

Chorusing means saying one sentence in unison with the recording, on a loop, until your voice disappears into the speaker's. It is narrow and repetitive, and it is the most precise of the three. Free audio editors such as Audacity will loop a selection for you.

Echoing adds a pause. You listen to a few words, wait a second while the phrase replays in your short-term memory, and then imitate that inner echo instead of your usual pronunciation. It is slower than shadowing and gives your ear time to notice what was actually said.

Kungfucius's suggested order for difficult material is echo first, chorus the stubborn sentences, and shadow last, once the clip is familiar.

Which polyglots and teachers recommend shadowing

The names below are often lumped together as "shadowing fans". What they recommend is not the same thing.

• Alexander Arguelles is the technique's best-known advocate. His version is specific: audio with no pauses, a bilingual text, and shadowing in a loud, clear voice while walking briskly outdoors with upright posture. He works in stages, first shadowing blind, then with the translation in hand, then with the target text. He has favoured Assimil courses as material. Passers-by are part of the experience.
• Luca Lampariello uses shadowing as one step among several. In his own three-month plan it comes fifth in a six-step routine for each lesson, after listening with the text, reading for meaning, listening without the text and reading aloud. His pronunciation course also warns that "blind shadowing" often fails, because you filter new sounds through your native ear, and that listening has to be trained first.
• Olle Kjellin, a Swedish physician and phonetician, is the source of chorusing, not shadowing. In his 1998 paper "Accent Addition" the teacher models a phrase, and learners then repeat it in chorus 50 to 100 times, rhythm and melody first, whole phrases instead of single words. He is candid that the approach had not been tested experimentally.
• Karen Steffen Chung of National Taiwan University teaches the Echo Method and is openly sceptical of shadowing for ordinary learners. Her argument, set out in a short article comparing the three approaches: shadowing was designed for interpreters who already speak well, and it leaves no time to listen closely. She recommends echoing short phrases, then listen-and-repeat, then reading in sync with the audio.
• Olle Linge of Hacking Chinese recommends mimicking native recordings, including looping a sentence in Audacity and recording yourself over the original to compare.

The common ground is narrower than the label suggests: imitate real speech closely, out loud, many times. They disagree about how fast and how blind.

How to practise shadowing for Chinese tones

Tones are where shadowing earns its place for Mandarin learners, with two conditions.

First, get the single tones and tone pairs reasonably right in isolation, with feedback. The guide to Chinese tones covers that stage. Shadowing a full sentence before then mostly rehearses your current mistakes at speed.

Second, treat the sentence as a melody. Tones in connected speech are shorter and flatter than in the textbook chart, third tones are mostly low, and neutral syllables nearly vanish. On your mumbling pass, hum the pitch contour without the words. Then add the syllables back.

Choose slow, clear audio at first, such as textbook dialogues or learner podcasts, and one speaker whose voice sits near your own range. For any sentence where your recording sounds wrong and you cannot say why, stop shadowing and chorus that sentence alone.

A first two weeks of shadowing

• Choose three clips of about a minute each, with transcripts, at a level you understand.
• Days 1 to 4: one clip, steps 1 to 6 above, ten minutes a day.
• Days 5 to 8: the second clip. Chorus the two hardest sentences of each clip.
• Days 9 to 12: the third clip, then return to the first and record it again.
• Compare the first and last recordings of clip one. If the rhythm has improved, keep the habit at ten minutes a day. If nothing has changed, the problem is probably in hearing the sounds, and time with a tutor will help more than another fortnight of chasing a recording.