What Is Sentence Mining?

Sentence mining is the practice of collecting real sentences that contain one new word (or phrase, or grammar point) and turning them into flashcards. Instead of memorizing meticulous = very careful, you save "She kept meticulous records of every expense" and learn the word inside a sentence you actually met. The method sits between reading for input and deliberate flashcard study.

Where sentence mining comes from

Using example sentences to learn words is old, but "sentence mining" as a named method grew up in online self-study communities in the 2000s, notably around the AJATT ("All Japanese All The Time") blog, where learners mined sentences from native media into the flashcard program Anki. The idea borrows Stephen Krashen's notion of comprehensible input at i+1: the best sentence to mine is one you understand except for a single new element. Sentence mining is a community practice rather than a formally studied method, but it rests on well-supported findings about context, retrieval and spacing.

A sentence mining example

Why it works for vocabulary

• Context carries meaning. The sentence shows collocations, grammar and tone, not just a translation.
• Personal relevance. Sentences from your own reading are easier to remember than textbook examples, because you remember where you met them.
• Retrieval and spacing. Reviewing the cards in a spaced repetition app adds the well-documented benefits of active recall and spacing (Roediger and Karpicke 2006; Cepeda et al. 2006).
• It bridges incidental and intentional learning. Words met once in reading are often forgotten. Mining the sentence turns a single incidental encounter into repeated, deliberate practice.

How to start sentence mining

1. Read or watch material you mostly understand. If every sentence has several unknown words, you will not find clean i+1 sentences.
2. Save words as you meet them. Mark the word and keep the sentence. Wordie lets you look up and save a word in one click while browsing, and export your saved words as CSV; you can then paste the source sentence next to each word in the spreadsheet before importing.
3. Be selective. Mine words that are useful or keep recurring, not every rare word you see.
4. Make the card. Put the sentence on the front with the target word bolded, or blanked out as a cloze deletion. Add a short definition on the back. The cloze deletion generator and flashcard CSV converter speed this up.
5. Import and review daily. Import into Anki or another app (see how to import words into Anki) and keep new cards to a sustainable 5–15 a day.

Common mistakes and limitations

• Mining too much. The bottleneck is reviews, not collection. Large backlogs lead to burnout.
• Sentences with several unknowns. These become hard, frustrating cards. Stick to one new item per card.
• Recognizing the sentence, not the word. Long sentences can be memorized as a whole. Keep them short, or use cloze cards so you must produce the word.
• Treating it as the whole plan. Sentence mining depends on a steady supply of input. Keep reading and listening; the mined cards only reinforce what that input brings in.

For a full workflow, see the sentence mining guide.

Sources: Krashen, S. (1985) The Input Hypothesis: Issues and Implications; Roediger, H.L. & Karpicke, J.D. (2006) Test-enhanced learning, Psychological Science; Cepeda, N.J. et al. (2006) Distributed practice in verbal recall tasks, Psychological Bulletin; Webb, S. (2007) The effects of repetition on vocabulary knowledge, Applied Linguistics.