Dec. 29 - Sentence Mining with Anki

in

Summary of MIA

The MIA curriculum for learning is simple in theory but complicated in practice. It breaks down into 3 steps:

  1. Actively immerse in content: Pay 100% attention to that content. Focus on the sounds of the language, then the words, then the meaning.

  2. Vocabulary study with an SRS: identify 10 new words per day. Take the full sentence of the word and make a flashcard in a spaced repetition system (Anki is the SRS of choice).

  3. Passively immerse with previous content: convert the active content into audio that you can listen to while cooking, cleaning, or walking.

Active Immersion

This one is the most straightforward. Simply log onto youtube, or Netflix, or Spotify and identify TV shows to watch and Podcasts to listen to.

Sentence Mining & Vocabulary Study

I struggled with this one. The founders of MIA talk about different ways to mine sentences. The basic method is just choosing 10 sentences a day out of your immersion and making cards for them manually. Then there are more complicated methods like audio cards, automated generation of flashcards, monolingual transition etc.

For me, I was looking for a way to do this sentence mining process while I’m on the go using my phone and removing the decision-making process from my normal immersion. Hence this blog post!

Passive Immersion

There are many programs that can extract the audio from video to make passive immersion easier. MIA also released a tool to help strip non-dialogue from passive immersion to make content even denser. There are already guides for this so I’m not going to include it here. If you want to know how to do this easily on Mac feel free to message me.

My Approach and Toolset for Sentence Mining

1) Choose Suitable Content

Find content that has native audio and native subtitles. Dubbed content usually has subtitles that don’t match the audio. Also, dubbed content is usually simpler than native content.

Some TV shows I’m watching:

  • La Reina Del Sur (Telenovela from Mexico)

  • Diablero (Fantasy series from Mexico)

  • Antes de Perder (Comedy series from Spain)

  • Millenials (Comedy series from Argentina)

2) Watch Content

Actively immerse with the content. Only watch with audio, not subtitles.

3) Download Content and Subtitles

This section is technically illegal so if you want to know what tools I’m using feel free to message me

4) Convert Content Into Flashcards

On Mac, I’m using the program SubStudy to automatically slice up the TV show into flashcards. Each flashcard includes the audio, an image from the scene, and the subtitle (previous, current, and next).

5) Import the flashcards into Anki and format

Front of card: audio and screenshot
Back of card: The dialogue with the current line highlighted blue, an audio replay button, and a collapsable section that we’ll fill in later with a definition

6) Go through the show line by line

I’ve configured the Anki deck to show 10 new cards per day with no review cards. Then I essentially rewatch the show line by line, first listening to the audio and trying to make sense of it. Then looking at the subtitle. If I understand everything in the subtitle, I will suspend the card. A suspended card doesn’t count towards the 10 new cards. If I don’t understand something, I will prep the card for future use.

7) Prep cards for future learning

For each thing in the subtitle that I don’t understand I begin the process of trying to understand it:

  1. Look up the word in a Spanish monolingual dictionary

  2. If I don’t understand, look up in Spanish-English dictionary

  3. Try to identify slang or idioms and use google to verify their meaning

  4. Fill in the explanation collapsable section with all relevant information. Try to use only Spanish in the explanation, preferably copy/pasted from the Spanish only dictionary

8) Move card to a review deck in Anki

I will confirm the card as learned in the first deck which will stop it from appearing again. Later I will move the card to a separate deck that contains all vocabulary sentences I have mined. This deck utilizes the default Anki settings.

Notes

This approach goes against the MIA philosophy of not being OCD and only sentence mining things that come up frequently, but the OCD approach is easier for me. It means I don’t have to make decisions about what is important to remember. It also consolidates the tools and lets me use Anki for mining and for SRS.