Moving from hesitant shadowing to keeping up with the audio feels like clear progress. But speed only tells you how closely your voice follows the recording. It does not automatically show comprehension, memory, or whether you can use the language in a new situation.
Use speed as one observation, not the entire grade. Also ask: Do I understand this? How much can I recall without the text? Can I use a chunk in a sentence of my own?
What a Japanese study measured separately
Ryotaro Hashizaki's 2021 J-STAGE study examined repetition, material difficulty and English chunk memory in paragraph shadowing. Participants shadowed two passages of different difficulty 30 times each, took cued-recall tests after every ten repetitions, and completed a delayed test one week later. Their shadowing speed was measured before and after training.
Recall of English chunks increased with repetition. The more difficult material required fewer repetitions in the reported comparison. Speed improved after training, but the author explicitly states that faster shadowing was not a prerequisite for memorising the chunks.
The tasks, texts and measures were specific. This does not establish a universal 30-repetition rule, prove that harder material is always better, or make speed useless. It gives us a practical distinction: measure fluency and recall separately.
Four passes that test more than synchronisation
Pass one: understand before speaking
Choose two to four sentences whose overall meaning you can follow. Check key words and mark meaningful chunks, such as *I'd appreciate it if…* or *the reason I'm calling is…*. Listen once for the speaker's purpose.
Pass two: gradually remove the text
Shadow with the full text, then with only keywords, and finally from a meaning prompt without looking. When a chunk disappears, revisit that chunk rather than turning every repetition into a word-by-word chase.
Pass three: test recall, not only timing
Stop the audio. Use the first word or a meaning cue to produce the complete target. If you can convey the gist but lose the key expression, practise that expression. Test again the next day and a week later; familiarity in the current session is not durable memory.
Pass four: increase speed only after stability
Once meaning and key chunks are secure, compare the fluency of your attempts. Speed should be a consequence of control, not something bought by swallowing endings or erasing stress.
Choosing difficulty without a magic repetition count
The finding about difficult material is not an instruction to choose a text full of unknown words. Recall depends on difficulty and existing knowledge, so there is no universally optimal count.
For self-study, choose a passage you mostly understand with a few useful new expressions. If you need a dictionary every third word, shorten it, use an easier version or prepare first. If it is already effortless, move on to adapting the chunks instead of mechanically completing thirty repetitions.
Taking the practice off the recording with MirrorSay
Save a short sentence you genuinely need in MirrorSay. Listen, shadow and inspect the timing once, then shift from matching the recording to producing the meaning. With paired bilingual material, speak from the meaning before revealing the English. Otherwise, cover the text and use keywords as a cue before replaying your recording.
Continue with Retrieval practice for speaking and Formulaic language chunks. To use the app feedback, see the illustrated shadowing guide.
Source
- Hashizaki, R. (2021). The Effect of Shadowing on Memorization of English Chunks and Its Theoretical Basis: Influence of Repetition and Material Difficulty. *LET Chubu Journal*, 31, 39–52. J-STAGE abstract and full-text access · DOI
This article is based on the public J-STAGE research record. The self-study routine is an application, not a replica of the original teaching procedure.