How to learn vocabulary (research deep dive)

The more deep knowledge a chanel has the fewer subscribers it has.

I appreciate you taking the time to create and share this video. Japanese learner here (hopefully Spanish one day), but my biggest issue is that I get ahead of myself and jump into content that is above my level. This makes me think that I haven't learned anything in three years, then I scale back, and the cycle continues. I think my expectations around language learning are unrealistic, even when I truly believe I am being realistic.

Thank you a lot, I really enjoy every episode of your podcast.

thank you for this amazing content! You answered some of my questions.

A fantastic video; very informative and practical!

Thanks for giving us these advices 🎉

I find that explicit vocabulary study is much less important for the initial few thousand words when first learning a new language because those are words that will come up extremely often. I would not take a frequency list and try to “kick-start” by attempting to memorize those - I would spend much more time learning naturally using graded readers in the beginning with maybe a bit of supplemental explicit study. I think this is more fun and interesting and you learn the words in a far richer and more complete way.

I agree with your suspicions on the studies that suggest the ineffectiveness of reading. I have spent much time learning vocabulary by making flash cards that test my recall of words in my target language. If I know a word, I can typically produce it without an issue. However as my desire to grow my vocabulary has taken me away from the core 5000 or so words , each word having their own flash card was becoming very time consuming. I turned to LingQ and was surprised how easily I could recognize/ recall words when I studied their flash cards. This was NOT the case from reading in a target language when I had a significantly lower vocabulary. I am also curious who the subjects were for this study. From my point of view, here in the United States, most “language learners” are typically concerned with passing a class that while getting the highest grade achievable, will be highly unable to communicate in their target language and the majority of the rest are relying on the newest App to have cracked the language learning code. Thanks for the content, I learn something great with each of your videos!

I developed my own way of studying vocabulary that I much prefer to flashcards with spaced repetition. It uses 3 copies of a text containing 20-25 vocabulary elements to be learned, and the vocabulary-to-be-learned is “played with” to practice both passively and actively over a series of sessions on different days (the spacing of which is under your own control). I can explain how I set things up but the advantages over flashcards are: it is faster to set up (and low-tech to boot -no software to learn), the vocabulary is studied with its full, meaningful context, it is very precise with only one right answer when you get to the active study step and finally, unlike with an SRS, your vocabulary study is fully under your own control (nothing is “due”, you study when you want to and for as long as you want to, you can prioritize the kind of language-learning you feel like doing that day whether vocabulary practice or other).

Useful summary! Seems to me that the key issue when using a mass sentence SRS approach is the sheer volume. Using your definition of fluency, we have at least 6000 words and collocations to master. Plus we have a number of grammar patterns. Plus in many languages we have a huge number of inflexions. If we put each of these on individual cards, it's all going to get out of hand. An idea I've been playing with, which I haven't really seen discussed, is progressive aggregation. I'd appreciate your feedback. Let's imagine someone learning English from scratch. They might start with something as simple as "the house". After 3-4 exposures they know it pretty well. And they might learn the verb "I leave" the same way. Now they want to learn the conjunction "and". So they can delete the first two cards and create a card for "I leave the house and...". Eventually this might evolve into "I left the house late and nearly missed the 9:00 bus". This card consolidates 5 or 6 more detailed cards that can all be deleted. And the learner starts to get comfortable producing more realistic complex sentences. If new words or patterns are added one at a time, the cognitive load should remain manageable. It's the mechanics of making this consolidation process efficient that I'm still playing with, but I'm already convinced that it can help tame the explosion of Anki reviews involved in your 1-year schedule of learning 20 words a day + grammar. And the process of combining cards becomes a useful review in itself. What do you think? Is there anything in this?

I understand when listen to someone teaching english even without subtitle, but when I'm watching video such national geographic or documentary for example,i really got stucked in my listening even my comprehension of reading on the trascript

Hi Loïs, I really enjoy the deep dive and was wondering if you are familiar with the literature (80's onward) measuring vocabulary size of native speakers. I skimmed through a few randomly (each with its own mini-survey) and the vast variation (by a factor of 10) left me confused. The inconsistent definition of a "word" (from any literal word to an overly restrictive lemma-based definition with random exceptions) made me realize that the numbers are pretty arbitrary. After reading the methodology of how they rate a participant as "knowing" a word, this sub-field seems ludicrous. Have you read through many of these (and can suggest a good one) and/or have your thoughts? I think that an Anki deck as used by a medium sized Reddit community would be more accurate than the handful of papers I've read.

Thanks

Really liked this video. When you make your anki cards from Netflix, are you including audio for the sentences and/or a screenshot? Do you think that is helpful or unnecessary?

Love you forever

LingQ anki migaku sentence mining

No way...! But In my case, as a person acquired Japanese (1 year and a half to the highest proficiency N1 from scratch) mainly through KZhead videos and light novel reading, I never ever deliberately tried to memorize any of a vocabulary or grammar... I don't think I'm a special case or something. I typically do only one thing probably special is that I looked up the dictionary at a frequent pace during my input consuming process. Now I‘m doing the same thing with my Korean and I’m kind of satisfied with my progress

Thanks for sharing the knowledge. one suggestion- It would be a lot nicer and bring to the next level in conversation if you can take out filler words, and limit hesitation.

A major problem that I've had with flashcards is that as the number of words you're learning is growing you are left with less and less time for new vocabulary until you actually have no time at all (because all your time is consumed by repeating the vocabulary that you have learned so far). I have tried to overcome this by retiring the vocabulary that is a few months old (that means I will no longer repeat that vocabulary), but it's not an ideal solution. Any ideas anyone?

So basically it's fair to summarize by saying that explicit vocabulary memorization is only useful paired with lots of listening, reading, and speaking, and lots of listening, reading, and speaking, without explicit vocabulary memorization on the side, is just way too inefficient. You'd need thousands of hours to reach a decent level (like people trying to use Refold, for example).