Hello just a quick note, with LinkQ you can export to Anki. So you would have both options to learn in LinkQ and also in Anki. No vocabulary needs to be created and you can start right away.
āThe huge problem appeared: I could not, for the life of me, make them stick, forgeting not just the next day, the next 5 minutes. The reason is, there is no context, nothing I can connect the word to.ā
I have gone through exactly the same experience with Russian.
However, it is inaccurate to say āthe reason isā¦ā.
What is more accurate is to say āone of the reasons isā¦ā
(long) explanation followsā¦
Weāll get your reason out of the way first (this is only ONE of the reasons, not THE reason), then weāll dig into some of the othersā¦
If youāve done anki or another SRS before with a language that is close to your own then you will have noticed that the words that are similar to ones from your native language are DEAD EASY to remember. They click very quickly. They are typically similar looking, similar sounding and with a similar or identical meaning.
There are, however, some words that are hard to remember. Typically they are further away or completely different sounding or have a different meaning.
These are non-cognates and are up to 10X as difficult to get them to āstickā as for cognates or close-to-cognate. That is the reason for that.
As you have correctly pointed out, Korean is a distant language from your native language. This means with no shadow of a doubt that most of the vocabulary is non-cognate, which makes them by definition more difficult to stick.
Compounding that could be different pronunciation of vowels and consonants, clusters of unfamiliar consonants combined with perhaps tones.
If you have all of those things together your brain will not even comprehend what itās hearing never mind be able to match it to something. Your brain needs to be able to recall something to make the link. If there is no representation you will need to build one by repeated exposure over and over and over and over over and over and over and over.
What this means, as you have correctly pointed out is that in the beginning āit just doesnāt stickā.
However. There is no way round this. Regardless of which method you use, you will have to learn from scratch the tongue placement, fricative sounds, palatal sounds, consonant clusters, tones and generally try to remember the gibberish.
It is not anki that is at fault, it is that the language is distant.
For reference: I had the exact same experience with Russian and even worse when I tried Mandarin for a month.
For French or Spanish I was able to get about an 85% retention rate on average for new words over the entire six month period I did anki.
For Russian, the first six months I got about a 30% retention rate for new words.
When I tried Mandarin for one month I got about an 8% retention rate for new words.
With some research into phonetics I eventually figured out that I was trying to layer on my understanding of English consonants and vowels onto Russian and English consonant clusters. It didnāt work.
My current issue is to optimize accumulating enough advanced vocabulary to be able to understand the more advanced content.
Logically low frequency words across the entire corpus of the language are far, far, far too low frequency to help out with more specific topics such as e.g. Friends episodes or gardening books.
That is typically handled by just watching tons of friends episodes (if friends is your thing) or reading tons of gardening books (if gardening is your thing).
Following the logic chain the question then becomes:
within those specific domains of language is there then a new set of higher frequency words that come up more often within those domains?
I think the answer is likely to be YES. We even have a word for it (in English). It is JARGON. There is gardening jargon and there is likely friends jargon.
In which case the option opens up for identifying the higher frequency words (jargon) within those domains and just memorizing THOSE.
So how to do that? Just watch everything over and over and read over and over. Seems much more time consuming than it has to be.
But how to identify the higher frequency words (jargon) that are only within those domains and are not otherwise high frequency?
I donāt have a good answer to that. Maybe going through the friends episodes and spitting out the ānewā words would be half a solution, but thereās no easy way to tell if itās repeated e.g. 9 times in the series (maybe worth learning, maybe possibly itās a jargon word) or just said once in the entire series (i.e. not jargon but lingQ has identified it as a new word).
I donāt know how to solve that.
YET.
Edit: I found this scientific paper for āde-jargonizingā TED talks.
Likely the tool they used has some kind of jargon detector in order to be able to de-jargonize. Therefore, being able to identify jargon would IMO potentially allow us to spit out the jargon for the particular domain (i.e. twilight books or how-i-met-your-mother episodes).
Automatic jargon identifier for scientists engaging with the public and science communication educators - PMC (nih.gov)
xttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC5549884/