×

We use cookies to help make LingQ better. By visiting the site, you agree to our cookie policy.

JBelly

gb   United Kingdom

Should I be using Anki in addition to LingQ?

May 20 at 07:02

I’m a huge LingQ supporter and have self learnt Arabic over the last 3 years. I’m at the level of being able to read Harry Potter and watch Netflix cartoons and pretty pleased with progress as the language is fairly challenging.

I’ve never tried Anki but see it get mentioned a lot on the forums. Until now I’ve ignored anything that takes me away from LingQ due to the sheer enjoyment I get here and like many others, I spent years ‘playing’ at language learning before I discovered LingQ so I’m hesitant to waste time again. However, it would be stupid of me not to enhance my learning if it could be beneficial.

All I know is that it is a flashcard system and can‘t comprehend how this could be better than LingQ SRS??

Also, I use my iPad here but a quick internet search says Anki is free on the PC but not IOS - is it worth me trying it on the PC?

Many thanks.

bankeinanin

us   United States
Wednesday at 20:41

I've been very off and on again about Anki. It really depends on one's level and it's never something anyone has to do. It can be a great accerant. Just like LingQ itself.

If I was new to a language, I would find a deck with native audio and sentences such as the Refold decks.

Maybe I would sentence mine with LingQ after that or with another app. Or I would just use Olly Richards books or just lingQ for my intensive study after that.

But at my level where my input level in Spanish is around B2/C1 and my output is about B1, I found a use for Anki that works for me.

I take every word and phrase that my italki teacher types out for me that was either new or I hadn't yet acquired fully, I make a lesson for that list of 30-50 phrases in LingQ and then after highlighting everything I make anki cards out of them. It has been very helpful in kinda filling in holes what I've acquired that I would use in everyday conversation.

At my level I don't find making anki cards from my regular LingQ lessons because most words I come across are pretty rare and I would find myself trying to beat some form of rare slang or a very rare word that's not that important to know through SRS.

Another way I "SRS" is that I randomly re-read a page in a book that I've already read through my comprehensive reading.

Anxxos

je   Jersey
May 28 at 16:31

@JBelly

Look at the length of the conversation you have started!

A Mexican army of divergent opinions, all passionately pointing into different directions...

Anyways, if you want to experiment with "proper" spaced repetition systems, here is a link where you'll find plenty of ready-made Arabic collections of flashcards. Maybe one will strike you as potentially helpful.

https://polyglotclub.com/wiki/Language/Multiple-languages/Culture/Helpful-Anki-Shared-Decks

JBelly

gb   United Kingdom
May 29 at 08:12

@Anxxos - I know, I’m surprised at the number of comments, in a good way. Thanks for the link - I’ll definitely have a browse!

victor866

es   Spain
May 28 at 10:23

I tried to use Anki long time ago and I think it’s a good method to learn and make stick words into your brain but is not for me.

I always forgot to do my reviews at the end it felt like too much work and I wasn’t enjoying practicing.

So even if it‘s a good method it needs to resonate with your way of learning languages.

Try it, there is a tons of pre-made decks done for a lot of languages, from various topics, or you can do your own decks based on the content you enjoy or want to practice.

PeterBormann

de   Germany
May 28 at 11:36

Hi Victor,

The problem is this:

If you expect that artificial SRS like Anki, Memrise, etc. have to be "fun", then you're toast from minute 1.

But even reading / listening to stories (Assimil, LingQ's Mini Stories, etc.) is usually not "fun" below a B1-B2 level.

Reading / listening becomes "fun" when you can digest the stuff you're really "interested" in (for me: AI, social sciences, history, etc.). That's the case from a B1 / B1-B2 level upwards.

It's the same in math, programming, competitive sports, and probably all practical skill acquisition processes: all beginnings are difficult...

The problem is nowadays that everything should be "super easy", "super fast", "super fun", etc.

alex1029

us   United States
Tuesday at 23:28

i agree its not fun necessarily. To me the fun is seeing new words and i have a system of only 10 new cards a day so i don't get bogged down. But 100% without anki I couldn't speak the language without it.

In comparison im learning italian and i can't remember any of the words when i try to speak and i know if i used anki for them i could have ok conversations by now,

PeterBormann

de   Germany
May 28 at 09:20

@nfera

"One issue which I've noticed in Italian is that I can recognise a word pretty much 100% of the time, but not necessary say it. I suspect you experience a similar thing. I can say half the word, or I partially can produce the word, but maybe I mix up the ending or the second half."

Yes, I have the same issue in L2s such as Dutch and Portuguese at the moment (but not in Japanese because they butcher the English loan words so much that the Japanese equivalent is often "unrecognizable". For ex.: "coffee" in English -> "kohii" in Japanese :-)).

"I do wonder to what extent shadowing would provide a similar effect. If your goal of Anki is L1/NL -> L2/TL in order to practise using the language/vocabulary, so you can later produce it, shadowing to some extent may provide a similar benefit."

As I'm using Anki only for LingQ vocabulary that I've read / listened to (sometimes many times), I tend to skip the L2 -> L1 completely because it's way too simple in all my L2s (sole exception: Japanese).

So what I want to train is "active" recall (= harder retrieval compared to the easier recognition operations) for testing purposes.

If I have enough time for my Anki reviews (which isn't always the case), I also try to modify my responses, i.e., I play around with different sentence structures. Sometimes I try to invent my own little mini story, too.

Anyway, what I want to get rid of is the "parrot stage".

Shadowing is, at least for my taste, too much parroting. That's ok for pronunciation / accent training, but IMO not good enough for retrieval operations.

In short, viable functional equivalents for L1->L2 retrieval operations based on Anki or other SRS (Memrise & Co) are:

  • self talk
  • journaling
  • written summaries of what I've read / listened to
  • written chats with native speakers / chatbots.

In the near future, more and more oral interactions with generative AIs as well.

Those are all practices were the focus is on "active recall" operations -

and that's not the case with shadowing.

Anxxos

je   Jersey
May 28 at 10:05

@PeterBormann

Many good points here.

Remember the structural weakness of shadowing output: it is devoid of morphology/syntax mistakes.

You want to produce your own output that commits those mistakes so you can discover them, tackle them and learn from them.

Making mistakes makes the own output generation painful: you thought you knew, but you don't, you thought you could, but you can't. Again and again.

However, the cognitive benefit is unparalleled and irreplaceable.

(By the way, this is where Krashen/Kaufmann perhaps place a little too much faith in their comprehensible input approach. It cannot suffice to make you a competent speaker for the reason above.)

PeterBormann

de   Germany
May 28 at 10:44

"You want to produce your own output that commits those mistakes so you can discover them, tackle them and learn from them."

Exactly: as learners, we want to have a kind of immediate feedback loop based on deliberate practice!

The British medical student who @xxdb mentioned and learned Mandarin super fast with almost native-like tones is Will Hart.

By using his Chinese girlfriend / friends, Chinese online content, and Anki, he was able to establish such a deliberate practice routine with great success.

See:

https://imlearningmandarin.com/2022/06/04/interviewing-this-master-of-oral-chinese-made-me-rethink-everything-i-believed-about-language-learning/

I'm trying to generalize this approach using

  • several thousand everyday dialogues (podcasts, Youtube, Netflix)
  • Deepl
  • LingQ
  • Anki
  • generative AIs à la ChatGPT

for various L2s.

Of course, this is not a "fun-first", but an "efficiency- and fluency-first" experiment so it's not for everybody. However, it's probably good to know what is possible in a certain SLA time frame - and what is not...

For me, the "generative AIs" are the game changers because they're like a plethora of L2 girlfriends/boyfriends on steroids who are never impatient, tired, etc.

Of course, you can't trust an AI blindly so there are definitely L2 structures that native speakers wouldn't use - depending on the specific use case. However, the hope is that with sufficient exposure to the L2, learners will sooner or later recognize most mistakes themselves...

PeterBormann

de   Germany
May 28 at 10:58

"By the way, this is where Krashen/Kaufmann perhaps place a little too much faith in their comprehensible input approach."

I'm not sure. But some seem to take the CI approach to extremes:

  • Grammar? No - comes naturally.
  • Pronunciation? No - comes naturally.
  • Speaking? Yes - after thousands of hours reading / listening first.

etc.

As an ex-language teacher, I prefer a more "flexible" learning strategy in this context that mixes different approaches depending on various factors (distance / closeness of the L2 to be acquired, SLA experience, time budget, money constraints, available tools, etc.).

In short, what works for person X in situation A will not necessarily work for person Z in a similar or different situation B.

Or as my Iranian colleague would say, "Life is simply messy, not well-ordered" :-)

Then it comes down to finding a "personal SLA mix" that is sustainable - and CI is just one approach among others...

nfera

au   Australia
May 28 at 10:51

@PeterBormann For me, I don't know if it's so much the 'active recall'/retrieval part. Like, when I want, I call recall that the cognate for 'complexity' is 'compl...ità'. It wasn't hard to 'find' this word, but it's just that 'recall' doesn't contain the complete word or the correct pronunciation, just like @Anxxos' 'panetteria' example. What if parroting is what will fill in the missing syllable?

@Anxxos "You want to produce your own output that commits those mistakes so you can discover them, tackle them and learn from them."

The issue is when you don't realise you're making grammatical mistakes. This is because no one corrects you. Or, alternatively, maybe the correcting does nothing. For instance, have you read 'The effect of adding supplementary writing to an extensive reading program' by Beniko Mason? An interesting paper.

https://www.researchgate.net/publication/296337703_The_effect_of_adding_supplementary_writing_to_an_extensive_reading_program

PeterBormann

de   Germany
May 28 at 11:14

"What if parroting is what will fill in the missing syllable?"

I'm not saying that we should "never parrot".

I mean we "all" do (in all of our languages) - more or less. That's why generative AIs can simulate our use of language: they are like "artificial parrots" (that don't understand meaning, but resort to statistics instead).

Or to paraphrase Roland Barthes:

We are born as originals, but we die as copies.

Unfortunately, as "pure originals" nobody understands us.

So, it's parrots all the way down: real parrots, human parrots, AI parrots :-)

And, maybe, even alien parrots if they are not xenomorphs that have no time for talking, just for hunting and reproducing...

To put it differently:

"Just experiment!"

Maybe it helps in your specific use case, but it probably won't help me in my L2s...

Anxxos

je   Jersey
May 28 at 15:40

@nfera

Thanks for the great Japanese study link.

Yep, correction without repetition will get you nowhere. Or at a glacial pace...

Error> flashcard > Drill, baby, drill

PeterBormann

de   Germany
May 29 at 09:42

@nfera

"Furthermore, two-thirds of the participants in English and Correction

groups said that they had copied from a book when they wrote a summary (Question 3), and a significant amount of the work written by the participants in both English summary and Correction groups was not their own (Question 4)." (Highlighting by me)

As I said:

"parrots all the way done" :-)

With this in mind, I find it hard to take the result of this study seriously because that's like saying:

"You can't lose weight by exercising! Well, I tried and hired a personal trainer who did most of the exercises instead of me! And I still lost almost no weight."

:-0

nfera

au   Australia
May 29 at 21:31

@PeterBormann It's true that a lot copied, but if you look at the amount they claimed to have copied, it's alright (obviously self-reported). The vast majority of writing was not copied. 33/38 and 30/34 participants claimed to have copied 25% or less. I'm not sure if this referred to the amount they copied in one particular writing piece or the amount of writing pieces they copied as a whole, but in any case, even if they only wrote 75% of what they should have on their own, this is still plenty of writing.

PeterBormann

de   Germany
May 30 at 05:03

@nfera

I agree that "grammar corrections" tend to work poorly.

However, a common experience in literary studies is that writing essays makes you a much better reader because it slows down the reading process considerably.

For example, simply reading more or less superficially a so-called "dark" poem by Arthur Rimbaud and writing in the L1 or L2 analytically about it (using advanced linguistic tools and specialized literature) is like "night"vs "day" or a "child" vs. an "educated adult".

In this sense, the reading + writing activities create "far superior", i.e., much more competent readers compared to the first group of "superficial readers".

I did this for many months at university (and I'm in the middle of doing it again reg. adv. communication / language research and generative AIs / NLP) and my experience was / is: the more linguistic / literary competence someone has (which includes writing!), the less dark the poems (or similar texts) become :-)

However, that's a different topic ("competence development reg. complex, esp. poetic texts").

What we discussed initially with the distinction "recognition vs recall / use / retrieval" (i.e.: L2->L1 operations vs. L1->L2 operations and their functional equivalents mentioned above) was rather:

Can recall / use / retrieval operations reduce the required number of recognition operations considerably?

  • This is more related to a "memory effect", sensu: "harder is better when it comes to memory (because the learners have to invest more mental energy)".
  • And that's also against a widespread "fun first - everything should be easy" attitude when it comes to effective and efficient (language) learning...

Another topic is a "feedback loop" where L2 learners want to have "corrections" in the SRS-based acquisition of collocations because there are several sources of error involved here:

  • translating 1:1 from L1 into the L2, which tends to create an L2 mess
  • creating L2 collocations / idioms by using L2 single words without knowing the former
  • using collocations / idioms that are simply wrong (often the grammar is ok, but the semantic nuances are off).

En bref, it doesn't make sense to put "wrong" collocations / idioms into our SRS and memorize them. And the general info principle that applies in this context is: "garbage in - garbage out".

In other words, a poor info / data quality should be avoided in both human cognition/communication/learning and machine learning....

nfera

au   Australia
May 30 at 08:36

@PeterBormann

Yeah, this experiment was only looking at the effect of supplementary writing on grammar accuracy. It was in relation to @Anxxos' comment about making mistakes to improve on them. From this, despite low sample size, you can at least be sceptical that writing to summarise, being corrected, and then re-writing the summary potentially has a negligible positive effect on accurate grammar production. Or, if it does, it's tiny and probably not worth the ~150 hours extra or writing and rewriting. I.e. writing + grammar corrections + rewriting probably doesn't improve your grammar. At least not in an efficient way.

I'm not saying writing has zero benefits, neither is this study saying that either. There are obvious benefits for writing: learning how to write (to be able to use that skill later, such as emailing), learning how to spell, practising recall/retrieval, etc. I disagree with the need to practise writing to be a 'more competent reader', because this seems like reinventing the wheel. A critical reader in German is likely to be a critical reader in Portuguese because, generally, it's an independent skill of which language you know (eg. when reading asking yourself the questions of 'What does the author mean here?', 'What is their cultural/historical background?' etc. no matter if it's German or Portuguese). But that's another discussion.

"Can recall / use / retrieval operations reduce the required number of recognition operations considerably?" [to be able to produce the word]

I think most people (input purists aside) believe the answer is yes. As mentioned before, I think my experience is that it takes X encounters to be able to recognise the word with a high confidence and then it takes another Y*X encounters to be able to produce that word, with Y > 1. That is, the number of encounters of a word through input to be able to produce the word is disproportionally larger than that to simply recognise the word. For this reason, as we all agree, simply more input is NOT the best solution to this problem. I.e. If you want to be able to speak/write (and can already understand), sitting on your arse and watching more anime is probably a dumb idea.

The issue that @Anxxos and I were mentioning might not necessarily be a 'partial recall', but rather a 'full/accurate recall' of a 'partial recognition'. That is, the brain only needs to remember 'pane..eria' = 'bakery', because in all cases encountered 'pane..eria' = 'bakery' is completely accurate and all that is required. Why remember the middle 'tt', when it's completely not necessary? Therefore, to remember the middle 'tt' requires that you make it necessary, as we are saying. There could be a number of ways for that: from matching pairs of 'What's missing in this spelling? Pane..eria' to writing to even listening (as listening many be slurred/not clear and therefore require it) to speaking (shadowing?), etc.

Hagowingchun

us   United States
May 28 at 04:33

@Jbelly Have you been mainly focusing on MSA or doing dialects because you said cartoons, but regardless assuming lingq works great with arabic like you can read all the words (i say this because of the arabic script leaving out short vowels) I would just keep using lingq if you are enjoying it and everything is going smooth. If you want to add a little bump to your learning add a little bit of anki, but I would just keep reading on lingq as long as the content is good (idk if arabic has a ton of importable stuff for lingq or not since its online pressence is less than say spanish etc.) I wish you the best of luck and keep on reading/listening!

Also I don't know how many resources exist for the dialects if that is scare than I would anki some of those words etc to get more exposure to them, but if dialect content is everywhere and plentiful I would just keep importing and moving along.

JBelly

gb   United Kingdom
May 28 at 10:01

@Hagowingchun - I’m learning MSA as this will provide me the basis for reading and watching the news as well as most subtitles are in MSA. If I ever choose to speak it then I will choose a dialect. As I’ve leant it from the beginning without vowels it is just normal for me. Some cartoons are in dialect with MSA subtitles so I’m unconsciously picking up dialects as I learn. The main reason for my question is that I am terrible at making my LingQs into known words. I have so many sitting on 3 and 4 as it is my crutch - I need to take the training wheels off and thought that a system like Anki or SRS many help convince me I do actually know these words. It’s silly really as I know I can always put them back down but we all have our quirks!

nfera

au   Australia
May 28 at 10:58

@Jbelly "The main reason for my question is that I am terrible at making my LingQs into known words. I have so many sitting on 3 and 4 as it is my crutch - I need to take the training wheels off and thought that a system like Anki or SRS many help convince me I do actually know these words. It’s silly really as I know I can always put them back down but we all have our quirks!"

Honestly, I don't think this is an issue. Whether they are marked as Known or marked as '3' or '4' doesn't really change anything. You are reading, you are listening, it doesn't matter what number you've given a particular word. Just move them up when you feel like it. When not, leaving it as a '4', isn't changing anything.

Hagowingchun

us   United States
May 28 at 12:51

@Jbelly nfera has a good point. I can relate to your situation like this with the status 3 situation. When I was learning spanish I did this with all cognates because I couldn't hear them in full native speech I felt that I didn't know them. Now my system is if I can hear it in the real world then I mark it as known (for cognates) and other words I mark known based on if I did'nt see this for a long time 6 months-9months/if it doesn't slow me down when i am reading at full speed then i mark. Everybody has their own system and the more time spent with lingq the more that system develops. I wish you the best of luck with your arabic adventures!

miriaml5

us   United States
May 28 at 03:56

Re: importing Lingq words to Anki. I tried and it was really cumbersome. Had thousands of words, not in any order really, and the sentences extracted were truncated —only a few words before and after the vocabulary word— so they didn’t always give the necessary context. I tried the “Morphman” extension on Anki which made it worse I think. Because it is supposed to order your cards so the most common words come up first. Somehow I had several cards all of the same sort of words. And they were all small words that don’t make sense out of context…

Keep in mind I’m not an Anki expert so I'm not the best at modifying it.

A time that Anki worked well I think is when I made sentence cards of a novel I was reading. Pick words you think are interesting/important to know, but that are medium frequency, enough that you would benefit from more exposure to them on Anki. I just made the cards myself and copied the sentence they came from as I was reading. That way I got the important context. And I think that did help me remember those words. I am reading through a novelist, and uses the words I made cards of over and over again. And I was able to remember them much more easily, probably because of the Anki, but it’s possible I would have remembered anyway. However, I got lazy about making sentence cards. Importing Lingq vocab into Anki is easier. But I couldn’t find a way to make it work for me well (tips welcome!) Pre-made Anki cards can be better but they are often from most common vocabulary. I stopped using Anki because I forgot and I’m fine with that, but I may pick up again. I think it’s also something to do if you’re too tired to read, but still want to work on your language.

alex1029

us   United States
May 27 at 16:53

yes, for things you think are important to remember or things you would use in conversation. Other than that things from books i dont think are super important to anki unless you really want to be natively fluent in reading ie not use a dictionary to get through a book

Paganono

de   Germany
May 27 at 15:52

Hi, many very gifted people like Luca Lampariello and Stephen Kaufmann don't recommend to use Anki or flashcards and I also made the experience, that it's sort of fun to learn vocabulary by heart (cause you may tell yourself " Wow I'm such an eager and perfect learner") but for me, in fact it doesn't really help to remember the words, when I see them in a text or want to speak. Therefore it's (for me) a waste of time.

mejiacapellan0814614

us   United States
May 27 at 13:42

I’m learning inglés now l want to learn

davideroccato

it   Italy
May 26 at 11:10

The example below just happened this morning and it's why LingQ (or this method) is way more powerful that "random" flashcards. (IMHO)

Plus connected to what @xxdb was saying about creating stories with generative AI à la ChatGPT for special groups of words and read them again in LingQ (although I would wait until reaching a certain number of known words in LingQ, unless there is a different need).

Right now, after having reached a certain number of known words in German (slowly but that's another story), my brain is more comfortable on doing also repetitions with the same subject without being extremely bored. (see mini-story boredom).

image

Every morning, instead of doing flashcards (that I used to do in the past) I do somehow something similar by repeating the same book I like over and over. When I finish the last lesson I restart from lesson 1 and I just do few minutes, not the entire lesson.

I read + TTS listening.

I focus only on the yellow words here and there and on listening, sentence structure and better comprehension of the story.

The story below is Moby Dick rewritten in simple and modern language.

In this story I still have more than 1000 yellow words!

Which means that if I keep going I will get those 1000 yellow words that are more difficult to my brain in this moment.

What happened this morning?

Well, I didn't know the word "Poller" (bollard) and it didn't stuck in my brain. Google images helped to have a visual reference but still didn't stuck.

BUT this morning, reading that page above and also visualizing the events that were happening on the boat, I have not only understood Poller but also I have understood all the words that you see in red. I probably watched in the past only once "geschlungen" or "genügte" or "Pfosten".

The fact is that all those words are connected to the story and they all helped to create the meaning of that initial word.

I know for sure one thing. Now I won't forget that word but I won't need a lot of repetitions for the others either. Because my brain "got" them now. They are not in the "difficult" or "not understandable" category anymore.

I definitely prefer doing this now as repetitions, in the morning, or night or when I go waiting to the doctor and in those "empty" moments.

It's way more fun and it works, and I believe using the same strategy with AI will boost a lot more than anything else.

IMHO.

Anxxos

je   Jersey
May 26 at 13:19

As persistence matters, enjoyment is critical. Kudos to you if you have found your way.

Never seen Poller before.

Quick check in my frequency dictionary: ranked 27,837.

Not exactly Grundwortschatz, but I'll take it, i.e. "flashcard" it.

- GPT, gimme 2 simple, concise sentences illustrative of the meaning!

- Sure. Here you go:

  • Der Poller am Hafen sichert Schiffe.
  • Die Stadt installierte Poller zum Schutz von Fußgängern.

I switch to Anki, always open on my desktop, click Add, create 2 flashcards. Done, will never forget Poller.

Different strokes for different folks.

Ogni testa è un mondo.

I have taught my kids that any new word is like an additional gold coin.

Thank you for the gold coin, !

davideroccato

it   Italy
May 26 at 14:12

Ogni testa è un mondo.

Funny thing, in a church not so far from where I live, on its bell tower, they have the banner: "Ognuno a modo suo!". :D

I have taught my kids that any new word is like an additional gold coin.

Wise teaching.

PeterBormann

de   Germany
May 28 at 08:02

"It's way more fun" (@Davide)

TBH, I doubt your choice of appropiate reading material at your language level:

1) Stories such as these have little to no connection to contemporary everyday life in German speaking countries, i.e.. the DACH region.

Therefore, vocabulary such as

  • Poller
  • Steuermann
  • Ausguck,

for example, is more or less useless.

And when learners then even use Google Search, ChatGPT, Anki, whatever to learn those words (in cotexts) by heart...

Sorry guys, there is nothing left of my mind to blow :-)

Yes, I know some LingQers can only read fantasy stories with dragons, dwarves, elves, etc. for SLA, otherwise they would be hiding under an SLA rock...

Well, if you want to "anki" a useful collocation in German, here it is:

"Schlimmer geht immer" :-)

2) And if the story does not have many "contemporary dialogues", it's even less useful.

3) And then you focus on "Imperfekt" forms such as "bot", "hielt", "genügte", etc. that are usually avoided in German conversations because the much simpler "Perfekt" forms (ich habe / du hast... + past participles) are used.

Reminds me of a similar problem in SW engineering.

Some coders focus on optimizing program routines (functions, methods, etc.) while their overall program designs and (choice of) algorithms are poor.. This wrong focus is usually a waste of time.

In short,

if you want to become fluent in an L2, then

* selecting the appropriate (reading / listening) material

* knowing what to learn and what to avoid (in the sense of "particular" fluency)

are crucial!

That is:

Focusing on apps (Anki, LingQ, this or that generative AI...) can't compensate for previous poor selection processes (similar to premature method optimizations compared to poor overall design / algorithmic decisions in SW programs)...

davideroccato

it   Italy
May 28 at 08:59

I actually agree but I have to deal with other issues (including many health ones that have the priority) and random material helped me to keep going. Between stopping, giving up and random or fun material I chose the latter.

I was thinking now the exact thing actually but strategy for German changed a lot in the last years so I went for "fun" or random to see what was going to happen.

However, this strategy helped also to understand many different things about the platform, including repetition that I do with books and TTS in the morning.

That might not be the best choice but yesterday night I was reading that one in the picture. It was nice to see that I could understand almost an entire story, and with all the difficulties I'm definitely happy with that.

There is theory and there is reality. Not an excuse though, I'm working on match them.

image

PeterBormann

de   Germany
May 28 at 10:06

Hi Davide,

"That might not be the best choice but yesterday night I was reading that one in the picture. It was nice to see that I could understand almost an entire story, and with all the difficulties I'm definitely happy with that."

Yes, they are awesome! I read hundreds of them when I was a kid.

Teachers (including my parents) tried to convince me that those comics wouldn't improve my reading comprehension in German - well, later I could show them that the woman who was responsible for the German texts was recognized for her sophisticated translations (E -> German) and wordplay.

En bref, Erika Fuchs is a German pop culture "legend":

https://de.wikipedia.org/wiki/Erika_Fuchs

BTW, I think comics are often excellent resources for learning everyday language (esp. slang)...

"helped me to keep going"

Well, if you really rely on established habits, you don't need motivational hacks (fun included) any more. Fun is just a nice-to-have, not a must-have.

I've come to the conclusion that "fun-first learners" have self-created problems that habit-based learners usually don't have - and esp. in tech / math domains you need a high habit-based "frustration tolerance", otherwise you're toast :-)

However, we discussed this topic ad nauseam in the past. Ergo, I simply wish you

Happy Pentecost holidays

Peter

PeterBormann

de   Germany
May 28 at 10:12

"There is theory and there is reality."

Yes. Or to paraphrase the all-time sage Mike Tyson:

We all have a theory until reality punches us in the face :-)

See, for example, my Chinese "hero" Xu Xiaodong

https://www.youtube.com/watch?v=Ycu7dvHBzk0

That said:

The fake Chinese kung fu masters were also "fun learners" :-0

nfera

au   Australia
May 28 at 11:21

@PeterBormann

"In short, if you want to become fluent in an L2, then

* selecting the appropriate (reading / listening) material

* knowing what to learn and what to avoid (in the sense of "particular" fluency)

are crucial!

That is: Focusing on apps (Anki, LingQ, this or that generative AI...) can't compensate for previous poor selection processes"

Well said.

I don't think every bit of material needs to be full of "comtemporary dialogues" though. Eventually, if you want to be able to read, you have to move onto reading material. Sooner or later, if you wish to read, you have to learn the tenses and vocabulary necessary for reading. This includes Präteritum in German and passato remoto in Italian, so you can eventually read the Adventures of Pinocchio.

As you well know though, I do agree that it's easier to acquire/learn the grammar and vocabulary of spoken language before that of books (it's often simplier grammar and vocabulary is more restricted), so, for most people, the progression which makes the most sense is first dialogue-heavy (including comics like Lustiges Taschenbuch), then that of books. But for those people who aren't interested in speaking the language nor listening to podcasts (eg. maybe only reading scientific research papers), then, in those cases, dialogue-heavy content is not relevant. But for most people, they do have the aim to understand spoken language, so dialogue-heavy content first would make more sense.

In relation to the rare, maritime words, maybe @davideroccato is a big fan of historical tall ships and his chief hobby is visiting maritime museums and participating in historical sea battle reenactments, for all I know. But, if not, it's probably not ideal wasting too much time on learning that word. But to each their own. Personally, I waste too much time reading the English dictionary, as I don't understand the word in English of the biology/anatomy/political jargon...

PeterBormann

de   Germany
May 28 at 12:45

"Eventually, if you want to be able to read, you have to move onto reading material."

You probably will be a better reader once you have thousands of everyday dialogues from YT videos, podcasts, and Netflix shows under your belt because these are already a few million words to be digested.

And as soon as learners read non-fiction and contemporary popular fiction, they have also a good foundation to read all kind of fiction (contemporary, non-contemporary, popular and non-popular).

On the other hand, it's simply a bad idea to rush forward and try to read, for example in French, Flaubert's "Madame Bovary" at an A2 / A2-B1 level where ca. 20-30 percent of the vocabulary is obsolete nowadays.

It's just a poor reading choice for beginners - often based on inexperience.

"But for those people who aren't interested in speaking the language nor listening to podcasts (eg. maybe only reading scientific research papers), then, in those cases, dialogue-heavy content is not relevant."

Yes. There are definitely cases where you need to be able to read scientific texts in Italian, for example, without being fluent in the language.

Nowadays, you'd probably simply use Deepl for this use case :-)

"In relation to the rare, maritime words, maybe @davideroccato"

Well, now it's too late: Davide has the potential to become a full-time observer of "Poller research".

And then he will publish YT vids en masse and discuss with Steve how reading Poller research texts on LingQ changed his life in an unprecedented way. "Poller research? - This is the way!" (I mean even The Mandalorian will appreciate a good old metal post in his spaceship).

"I waste too much time reading the English dictionary"

Same here with German.

It's always good to remember that we haven't even fully mastered our own L1s yet :-)

xxdb

gb   United Kingdom
May 27 at 00:40

I agree. For me, a couple more months in and I should have created stories with *all* of the remaining difficult words out of the original 10,000 head words. At that point it won't be necessary to create any more stories as I can just read them over and over like mini-stories. So yeah, good plan.

PeterBormann

de   Germany
May 28 at 06:00

Nice, Davide!

A German word that I normally neither use nor care about - because it's more of a "terminus technicus" (-> ships).

Many Germans would probably just use the general expression "Metallpfosten" - even related to pedestrian zones...

Be that as it may:

There is even a "Pollerforschungsbuch" where you can read 440

pages about all the things you never wanted to know about "Pollers":

https://adocs.de/de/buecher/design-raum/pollerforschung

I'd never learn or "anki" such words because they're usually not part of everyday language, but ok: "jedem Tierchen sein Plaisirchen" :-)

However, there's a general lesson to be learned here:

There is no such thing as "general fluency", only "specific / particular fluency" even in our L1s. Ergo, you should never ask me anything about the intricacies of whalers, whaling or whales :-)

davideroccato

it   Italy
May 28 at 09:34

There is even a "Pollerforschungsbuch" where you can read 440 pages

That is insane! 440 pages? Now I will never forget this "useless" word but I can have an happier sunday.

The spirit and proud history of the word Poller is working its way back to become relevant to the society again. One day, when I will be very old, I will tell this story to a group of kids that are approaching their first foreign language.

A very old me with a 3D Visor :

"Never underestimate the preciousness of an old word like Poller. As @Anxxos said long time ago, each word is a gold coin, even if we don't really understand the purpose. Basically like each person in the planet.

But also be careful, as @Peter said, don't waste your time with too many useful Pollers just for fun or you will be an idiot! Don't hang out with strangers you don't know or doubtful friends.

Be productive but also mindful. Learn to understand what you need but also go with the flow of life.

Old words are never really useless, they have a story, a past, a connection with the life and culture you are learning. Thousands, millions of words forgotten are waiting for you to be discovered again. They are happy to have a new life, even for just a moment.

The past will help you understand the present and in the present you will build the future."

Kids: "Hmmmmm! You mean like Metallpfosten?

Old me: "Yeah, don't worry about it, we fly now and build different things...".

Random Möwe in the future: "We still use them!"

image

PeterBormann

de   Germany
May 28 at 12:05

Ha ha ha, awesome!

But my scenario for the kids in the future is a bit different:

AI: "Hey, kids. Do you want to know how I became a "Pollerforschung"-Forschungs-AI?"

Kids: "You mean there are

* words

* useless words

* the research of useless words

* and then the AI-observation of the research of useless words?

First, can you estimate how many of our brain cells will die listening to this story?"

AI: "Kids today have no respect for older AIs!"

hiptothehop

us   United States
May 25 at 02:18

Back in the Pre-Anki days (1990s) I made a lot of flash cards for Japanese, and had my own manual SRS. My anecdotal impression is that the words stick best in my memory when it is not from a common words list or something, but from input or a conversation I just had. Just learning the next kanji on the 5th grade kanji list, did not seem to work as well as learning the kanji in words I experienced. There may be some cognitive bias here, but I feel that there is something connected to experiencing the word in a meaningful way and then recalling that when I review the words over the next few days that made the words "stick", almost like they were snapping into place in my mind.

Because of that, and because making cards is so time consuming, I no longer make or use any flashcards outside of LingQ. I only use the LingQ SRS system and nearly always sort by date created. If they fall off the list I don't sweat it. I just rely on input and conversation to give the words and phrases their proper frequency in my review.

davideroccato

it   Italy
May 25 at 10:04

Yes, from what they were talking below we can see that cards work if they are in sync with what we are reading or doing.

As I learnt my previous languages by immersion in the countries, what I did, without realising it, it's exacting something like this but by conversational production.

Basically, when people would tell me something, like an expression or new words and so on, I would repeat those new words as soon as I could. For example using them the next day with other people or in other occasions. I didn't force myself to do that, it's just a method that was natural to me. I wasn't embarrassed to make mistakes.

It seems this would fix in my mind all those words very fast giving my brain the flexibility to incorporate others very quickly.

This is making me rethinking the way I study now.

Anxxos

je   Jersey
May 25 at 10:09
@Anxxos (Sorry, I can't reply directly) Taking your "panetteria" example, do you expect it is worth the effort to learn/activate such words learn L1>L2 with repetition on Anki? Or, is the experience of not being able to produce it correctly once, and then learning it at that time not sufficient, and perhaps more efficient?
I may be too comfortable with my own ignorance or making mistakes :-), but it seems like having someone correct me or ask them to remind me of the word for "bread store" in the L2 at exactly the moment I need it, could be more efficient than spend time drilling with Anki L1>L2?

If you have access to a kind, altruistic person that will patiently, repeatedly provide hints and correct your mistakes on the fly, yeah, sure, that is ideal. There are many social contexts where that is rather unlikely, though.

Flashcards are a grind. I am not going to argue otherwise, but it will mechanically accelerate your gains, whatever else you do to progress. I see it as a polite way to return the favour to my patient interlocutors whose language I am butchering.

Properly used, they are so impressive.

I used them with my 3 kids. Their flashcard collections reached 100,000+ items over the years. Everything went in: math, languages, history, biology, art, mum and dad's phone numbers, etc... The recall was 90%! Imagine that! The average daily drill duration was under 10 minutes. (SRS used: Supermemo)

If you want to go straight to flashcard hell, easy: do single word cards and bite more than you can chew. A great masochistic experience. Guaranteed.

Learn slowly, but by rote, simple, frequent sentences and something strange will happen to your brain: it will develop a "cognitive village" of the foreign language in your head where the sentences will start to cross-pollinate. In my experience, it takes 3-4 months. And you're done. You're functional in the target language. Far from perfect of course, but functional.

Mileage will vary. Personality will vary. Optimal strategies will differ, but it is where I am after 20 years dabbling in those issues with great curiosity.

hiptothehop

us   United States
May 25 at 15:23

I very much appreciate this reply. Thank you.

You mention 3-4 months as a timeframe for the cross pollination to occur. At what point in the language learning process do you implement this? From the start? Do you continue through more advanced levels?

Thanks again.

Anxxos

je   Jersey
May 25 at 17:21

From Day 1.

The simplest sentences to start with.

Both ways always: L2>L1 (comes easy) and L1>L2 (comes hard).

Avoid more than 1 difficulty per card.

Difficulty is word, collocation, idiom, grammar point or syntax (any word order that feels unnatural to you).

Sentences teach you syntax. Which is critical.

If a card proves refractory to memorization, ask GPT to suggest another 3 flashcards around the same word or collocation or syntax, whatever the issue is.

Your brain will yield.

However, remember: it is about a process of biological engraving. The only evidence of integrity of the proceedings is correct, near-instant recall. My kids had 6 seconds to provide the right answer. No if, no but.

Calibrate knowing that if you add 10 new cards/day, you will end up with 4-5 times that as a number of daily revisions. It is important to start slow, to know your limits.

Another tip is to review your cards early in the day (pre-breakfast for me) so that you do not have that discouraging, Damocles sword of the dreaded drill-to-do above your head all day long. Works like a charm. Habit is a strong master.

All the way to advanced level. I use it every day for the 4 foreign languages that I continuously learn (plus many other things).

hiptothehop

us   United States
May 25 at 21:35

I very much appreciate the depth of this response. You've convinced me to try this with areas of Spanish (I'm about B2 or so) where my ability to understand has well outpaced my ability to produce. (double object pronouns with some verbs, some prepositions, some subjunctive constructions, etc.). I may use this in preparation for conversations on certain topics as well, in addition to reading articles.

Thanks again.

PeterBormann

de   Germany
May 28 at 07:03

"it takes 3-4 months. And you're done. You're functional in the target language."

Hm, if you do that for 1 h a day, you have ca. 90 h after 3 months.

In L2s that are distant from your L1 (let's say an English native speaker learning another tonal language such as Mandarin, Vietnamese, etc. or Japanese with its pitch accent), that's so little exposure that you're still a (n absolute) beginner.

And even in closer L2s, you're only at a parrot stage knowing set phrases (in the best case, that's B1, usually it's rather A2-B1).

What you can have after ca. 90 h - starting from scratch - are simple and slow conversations Benny Lewis style, but that's it: You are neither "done" nor "functional" (at least if you want to survive more free-flowing and fast-paced conversations with native speakers. And we're not even talking about writing here...).

I'm definitely pro "specific" Anki use cases, but your scenario is quite unrealistic...

Anxxos

je   Jersey
May 28 at 16:20

15 years ago, I was given 4 months to become "business functional" in Spanish (speaking French (N), English and German).

With 2x 90 minutes of weekly tutoring. Rather long and draining as I quickly had to do most of the talking.

At the end of the 4 months, I asked the teacher what my level was.

She said: "Solid B2. I have never seen that progress before. So, what is the full story here? You lived in Spain as a young kid and it is all coming back to you?"

Made me smile...

Unbeknownst to her, I had decided to use a SRS (Supermemo, which I still consider the best) in a totalitarian fashion. Every language data point about Spanish, I would "flashcard". Maniacally. Every word, every collocation, every grammar point, every odd syntax, every correction, everything.

Drill, baby, drill. Maybe 300-400 cards revised per day. 45-ish minutes a day.

With something like an hour of daily comprehensible input on the side (Lingq unknown to me at the time). My favourite were animal documentaries. Don't ask.

Worked like a charm.

xxdb

gb   United Kingdom
May 24 at 23:48

OK so I'm going to spitball a hypothesis here...

What IF you can speed up your memorization of active production by doing a simulated version of what English dude learning Chinese (Will something is his name)...

What he did was he immersed by having tons of Chinese friends and talking/studying/doing anki for 8 hours a day every day for a year. One of the things he did is if he didn't know a word in a conversation or a phrase he put it into anki and drilled it.

So... one thing I try to do (not often because I feel like I'm not really ready to speak consistenly yet)... anyhow... one thing I try to do is have a conversation with myself by doing both parts of the conversation.... I speak in russian to google translate and see if it translates to english correctly then I type the response in english (as if I'm the other person) and let it speak back to me in russian. So....

I find there are some words don't easily come to mind when I'm trying to speak. The handful of times I've tried this I have put the words into "learn again tomorrow" in anki.

But maybe creating a second deck with those words but instead of learning it as T2-T1, put it to T1-T2 so I see the english word which will force me to try to recall the russian word. But only for those I'm having trouble with. That should in theory speed up active speaking if the hypothesis is right.

PeterBormann

de   Germany
May 28 at 08:36

Sure, all language "use" helps:

  • self talk (e.g., summarizing what you've read / listened to)
  • journaling
  • using the phrases / collocations / idioms in chats - nowadays not only with persons, but also with chatbots
  • rewriting stuff using something similar to "Deepl Write"

etc.

This recommendation isn't new - only the generative AIs are the new app kids on the block.

Probably, we'll see that more often as soon as oral user interfaces are widely available for interacting with the generative AIs...

xxdb

gb   United Kingdom
May 24 at 11:22

@nfera:

My thinking was to of how to shortcut the required number of read while listens to be able to produce the word. I find if I can already recognise the word, then I say it out loud idk on 5 separate occassions, I can then independently produce the word. So I'm pondering if a way to shortcut this...

I love these kinds of discussions so I can't resist.

I surmise that your memory of a word includes a bunch of different components all connected to your internal representation of the word-concept. It includes the overall sequences of phonemes, the phonemes themselves, the situations in which you've heard the particular word, the combinations you've heard the word in, the different variations of the word AND how the phonemes *feel* in your mouth as you say them.

It is *possible* and I'm armchair quarterbacking here that one of the ways your brain knows that a word is important to be able to remember is that you are actually speaking it. IF that hypothesis is true then pronouncing the word should help you activate it and remember it. I can say that for myself I have "scenes" in my memory attached to words where I have tried to learn them by interacting with a real person. I also usually have an intense feeling of "I need to learn this word" if it's a word that should be easy in a conversation and I can't produce it. So yeah I do reckon anecdotally that speaking words should help cement them in place.

nfera

au   Australia
May 24 at 11:48

@xxdb Exactly! That's my thinking. All those input purists completely ignore that production actually strengthens your knowledge of a word (perhaps disproportionately). The ideal situation for strengthening your memory of the word, and eventually being able to produce it would be to have conversations with it, as you have the 'stress' to tell your brain 'this is important', like you're saying. Next down the list would be what @PeterBormann does with his L1/NL -> L2/TL flashcards (due to convenience and time reasons). My thought is that next would be merely repeating/parroting what others are saying, as in the case of shadowing (but you get the added benefit of pronunciation practice and a large amount of words spoken).

davideroccato

it   Italy
May 24 at 13:04

All those input purists completely ignore that production actually strengthens your knowledge of a word (perhaps disproportionately). The ideal situation for strengthening your memory of the word, and eventually being able to produce it would be to have conversations with it, as you have the 'stress' to tell your brain 'this is important', like you're saying.

I find this interesting connected to our previous discussions.

This last language I have been learning is the first language I study completely at home. Unfortunately I have also had too many health issues in the middle so I can't have an accurate experience to compare with a "normal" learning situation.

BUT I have learnt all my previous languages directly in the country: Ecuador, France and England. Without almost any previous knowledge of the language.

I wanted to do the same with Germany but many things have changed.

I have never enjoyed "learning" but using the language. With the knowledge I have today I can see that I should have done a lot more studying when I could so to upgrade my languages a lot more.

What you wrote was probably the ideal situation you were talking about. And maybe the fact that I don't experience some of the things some of you are experiencing comes from that type of mind conditioning.

My method was to learn by interacting as much as I could with anybody I could (often or always for free). A few grammar books here and there (some Anki, Assimil etc.) but mostly interacting with people and then working directly in the country.

For example, I couldn't survive in France the first days I went there with my car. I relied on a shop where the owner was from Uruguay and I asked her many things in Spanish to understand what to do. Step by step I learnt the language and then started to work. When I passed the C1 was very easy and the commision asked me why I didn't go for a C2. I said them I couldn't because it wasn't allowed if I didn't have another certificate before, the closest was the C1.

That would be the exact scenario where your brain doesn't have a choice. Either you learn and understand or you don't.

Which also means producing a lot, creating lots of brain association, and many more thing. To me is the best and easiest way to learn a language.

I didn't need to repeat a lot of times words or phrases, I used to use them immediately. What I didn't do is to constantly increasing and improving the language instead of relying on what I needed for living only.

Today I would do the same but adding reading with LingQ + grammar + more writing + don't waste precious time with nonsense activities!

It seems my brain capacity to remember the language is a lot stronger than doing countless repetitions. I've lost a lot because for many years I haven't used anything at all but I still remember an incredible quantity of things. And maybe, visualization of the words improved as well because for me it was "important" to speak well and don't mix the languages. So my "brain" learnt to visualize the words to avoid mixing them. Or at least, I analyse it that way.

I'm just thinking out loud to share but it could be that full immersion created those patterns in my brain that maybe is more difficult to create when learning at home. If we were able to force the same scenario giving the brain no choice in order to stimulate those survival patterns, probably things would be "easier".

xxdb

gb   United Kingdom
May 24 at 23:40

I do believe that some kind of immersion (whether simulated or real) is needed to level up to high intermediate or advanced. I don't know how to effectively simulate that without actually spending time in a target language environment. Which is the crux of the quest.

davideroccato

it   Italy
May 25 at 10:09

Yes, the problem is money as well. You could pay a lot of different tutors, daily, but it's going to be expensive. But I don't even know if it's the same thing honestly. It's different when you need to open a bank account, or going to different offices, or being forced to use the telephone to discuss a contract and so on. Or going to the open market and talk with the farmer. A lot of different situations and interactions.

nfera

au   Australia
May 25 at 10:42

@davideroccato With regard to those specific situations, you'll probably lack precise phrases and vocabulary, but it's not too much of an issue. You could get to an advanced level of reading literature and understanding everything in movies and still lack a tiny amount of specific vocabulary for such situations (as they don't appear in movies and books), but you'd still be fluent. Honestly, I don't think it's an issue. The first time to do something new in a different country is always a challenging experience, because you don't know the process. That would be the larger challenge than the language itself in those situations.

jt23

us   United States
May 25 at 10:17

@nfera My first tutoring session was quite difficult, so I decided to bear down on listening comprehension and speaking practice,

In sentence mode I now listen/repeat/shadow/repeat (not looking at text) 10-12x, while visualizing the action. By the last repetitions I have practically memorized the sentence and mostly nailed the pronunciation.

If the sentence is long or multiple, I break it down into clauses or phrases, then back-chain (start at the end) until I can do the whole passage.

This is slow. It can take a couple minutes to get through one sentence of Harry Potter. But I find it weirdly satisfying and I can check off all the New Words as Known.

PS. I lose interest if I try this with podcasts and mini-stories.

nfera

au   Australia
May 25 at 11:18

@jt23 Do you find this shadowing practise to help you with the problem @Anxxos and I were mentioning? That is, only partially being able to produce the word.

jt23

us   United States
May 25 at 15:50

@nfera I've only been doing this listen/repeat/shadow/speak (without text) routine for three weeks.

I find it has tightened my listening comprehension, pronunciation and ability to speak full-speed quite a bit.

It's also sharpened my sense of grammar and gender. I do remember the words when I see them again.

However, I can't say yet how well it works to activate my reading vocabulary for new output (conversation).

Arguably the sentences I'm using are too complex for the cross-pollination Anxxos describes and I don't drill on them past my first encounter.

suzmax

mx   Mexico
May 25 at 22:27

@jt23 In sentence mode I now listen/repeat/shadow/repeat (not looking at text) 10-12x, while visualizing the action. By the last repetitions I have practically memorized the sentence and mostly nailed the pronunciation.

I LOVE this idea! I'm starting it now.

hiptothehop

us   United States
May 25 at 02:35

@xxdb I can say that for myself I have "scenes" in my memory attached to words where I have tried to learn them by interacting with a real person. I also usually have an intense feeling of "I need to learn this word"...

Yes, I relate very much to this feeling of "scenes" being attached in my memory of when I first encounter, learn or use certain words. I sometimes remember that scene years after in relation to the word.

xxdb

gb   United Kingdom
May 25 at 09:41

Right? There are a half dozen "scenes" associated with certain words I learned in Spanish more than 20 years ago I can still remember clearly.

Here is one: "no entiendes cuando te digo 'se fueron' verdad? que tal si te digo 'han ido?'"

Translation: you don't understand when I ask "did they go away?". What if I ask "have they gone?"

Through sheer embarrassment I went directly to a book store and looked it up in a spanish dictionary. (The internet was crappy in those day).

xxdb

gb   United Kingdom
May 23 at 23:52

@nfera: (sorry I couldn't reply to your post directly)

I mentioned it on your other thread, but I think this could've been because your LingQ reading material was children's fantasy novels (like Treasure Island). If your LingQ material had of been conversations or graded readers (at the appropriate at which you were up to on your frequency list in Anki, eg. word idk 5,000), you would've had a different experience.

The graded readers and your Anki cards would be using a very similar frequency list, so you would've been seeing similar words.

Agreed. I was, however, unable to locate graded readers in my target TL. As I explained, the next best thing to that was trying to find children's books that were at my appropriate reading level (which I'm guessing is equivalent of the 3rd grade - approx 7-8 year old). As we've discussed, it didn't work however. I was training two separate things and didn't get synergy. I got some progress (i.e. I can now *read*) but I didn't get the progress I wanted because not enough overlap.

But, as your experience highlights, it's definitely possible to do two separate things, which are out of sync and don't complement each other much.

Correct. The synergy was lost because the target vocabulary wasn't matching.

That is, drilling high- and -mid-frequency words on Anki while doing intensive reading on fiction books (i.e. with many low-frequency words).

That however is not in fact it. I've analyzed the books. While you're correct that they do have low frequency words, my target vocabulary is now in the 7000+ headword frequency range. It's no longer mid-frequency. It's low frequency or low-mid-frequency. So it was quite simply that the low frequency words I was trying to learn were a mismatch for the ones in the book. [The book had neither the low frequency words I was interested in nor was I interested in the *different* low frequency words that it *did* have) so lose-lose.

Anyhow, while I agree with you about graded readers, I'm not sure there is any easy match for e.g. 7000-10,000 word frequency in graded readers which is the range needed to get you over the "finish line" to B2/C1. (Approx equivalent to 25K-35K in lingQ). That said, I have the solution now with chatGPT derived stories; artificial graded readers targeted precisely to me, which is basically a long winded way of saying I'm now doing what you suggested.

S.I.

ru   Russian Federation
May 23 at 16:03

No, you shouldn't waste your precious time on something so boring. Besides, all Anki users end up taking longer and longer bowel-movements, because the amount of cards and decks never gets smaller, it only grows forever.

And don't listen to those LingQ users who use this bloody tool. And don't try to help them, it's too late.

:>

xxdb

gb   United Kingdom
May 23 at 16:48

Thanks brother. That was hilarious.

image

lq  
May 23 at 12:39

I use ANKI but sometimes I feel like I'm married to it. I added way too many decks to my pile and has taken me up to 30 mins/day. I would say it works, but I also think LingQ comprehensible input gives you natural SRS anyway. I don't really get the way the LingQ SRS system works. ANKI is now telling me that I'm going to be doing these words for another 2 years. Like I say, it does work, but can feel endless.

xxdb

gb   United Kingdom
May 23 at 12:52

Yeah, it can get overwhelming if you take on too much. One particular trap you can get into is that it's EPIC for close-to-english languages. You can literally memorize 50-100 words a day using it for e.g. French or Spanish.

Try doing that with a language you don't even have the phonemes for and you will rapidly end up with 1,500 reps a day with no end in sight.

There are definitely pros and cons.

It is "efficient" but it is NOT fun.

If you cannot maintain the grind even when it gets brutal it's not for you.

So yeah. In theory it is the optimal way of memorizing the maximum number of words with the least daily effort. Which is not to say it is five minutes a day. It's still measured in half-an-hour to a couple hours.

It will therefore take away from your lingQ time. And you have to balance that with the point that language acquisition is not only just about the words you know, which is what anki is best for (vocabulary acquisition). You can do nothing without words but you can't talk without phrases and grammar. LingQ will give you everything.

PS: You *can* use anki to memorize phrases but I personally prefer to use lingQ for phrases because it's fun to read.

xxdb

gb   United Kingdom
May 23 at 12:34

I'm a huge anki fan (and probably one of the few people who are *honest* about their progress).

Anki is useful to get up and running quickly from zero for the most high frequency words (e.g. up till the first 1,000 or 2,000).

Beyond that your mileage may vary. I can make arguments pro and con for different techniques beyond 2,000 words but I'm not going to hijack this thread.

At the end of the day the most important thing is actually doing something *every single day*. That's what will get you success. The rest is just tweaks.

nfera

au   Australia
May 23 at 11:06

@JBelly If the goal is to increase your vocabulary, honestly, I wouldn't bother. Just keep doing what you are doing on LingQ. Reading (while listening) to content with New Words in domains where you want to acquire vocabulary.

asad100101

pk   Pakistan
May 21 at 11:20

I have a question for long/experienced users of Anki. Did you notice any considerable effects on the active vocabulary in your target languages? Did it increase? Did it increase by a small margin? If the margin is not noticeable when it comes to increased active vocabulary vs reading, does it make sense to spend time adding individual words/sentences in Anki? The better approach is to read more?

I passed official reading exams (texts were written in formal German, and had a higher word density). I scored 100% At A2 and 75% at B1 level. With no special prior reading practice. Just used LingQ for reading purposes and sat for the exam.

One thing is for sure, LingQ helped me with my passive vocabulary hence my results in reading tests.

To make it short, does Anki help with my active vocabulary or not? or is it for increasing passive vocabulary? If this is the case, then LingQ is doing a fantastic job as well.

PeterBormann

de   Germany
May 21 at 13:44

Hi Asad,

First, forget about the distinction "active / passive" that is related to the technical sender - receiver / input - output model of communication as data exchange where

sending of data (outputting) = seen as active

receiving of data (inputting) = seen as passive

However, that's not how "human" cognition and communication processes work...

In short, reading and listening are "always ongoing" operations that require the processing of prototypes, comparing competing language forms, resorting to background knowledge, etc. - there's nothing "passive" about them.

Second, a better distinction is probably "recognition" vs. "use" of language forms - and both are always ongoing operations. If there's no operation at all, nothing happens: no listening, no speaking, no reading, no writing - and in the extreme cases (i.e., severe coma or death): there are neither cognition nor communication.

Third, one of the main differences between recognizing and using language forms seems to be the different degree of mental effort being involved. That is, pure recognition requires less mental effort than the use = retrieval of language forms.

Fourth, against this background, I'd answer your question as follows:

"To make it short, does Anki help with my active vocabulary or not? or is it for increasing passive vocabulary? If this is the case, then LingQ is doing a fantastic job as well"

  • If you use Anki only for recognition purposes, i.e., L2->L1 operations, it's probably similar to reading and / while listening a lot. And in this case, LingQ is often sufficient, esp. at the lower / upper IM stages (however, ASRS are still useful for absolute beginners to get a boost and for ADV learners for learning less frequent collocations).
  • If you use Anki for L1->L2 retrieval operations, it improves the use of the L2 (your so called "active vocabulary") as well. And that's not always (at least "directly") the case with reading and / or listening alone because the recognition operations only tend to act as a mechanism to facilitate the use of language...

Fifth, the next question that @nfera asked me a few weeks ago, is this:

Why not skip L1->L2 retrieval operations using ASRS and resort directly to speaking and writing activities = language use?

Yes, that's "the" question in the age of generative AIs à la ChatGPT and Co that seem to get better by the minute.

For me, it's just a matter of convenience / time efficiency at the moment:

Since LingQ creates the Anki flashcards for me from L2 material I've listened to/read, I can immediately use Anki on the smartphone to test myself. I don't need any AI support for that.

However, generative AIs are useful to create more sentence variants for specific vocabulary that can be imported into LingQ.

That said:

I'm not sure what the AI future brings (and we have nothing seen yet!), but the better the AIs get, the less likely we may need ASRS such as Anki for L1->L2 retrieval operations because we can use the language forms directly in oral and / or written interactions with the AIs... (note: I'm referring to a more "seamless" interface between AI and user here!).

Schönen Sonntag

Peter

davideroccato

it   Italy
May 21 at 14:19

...one of the main differences between recognizing and using language forms seems to be the different degree of mental effort being involved. That is, pure recognition requires less mental effort than the use = retrieval of language forms.

I agree, and I could add that our mind is smart enough to save energy and skip the effort. With brain fog, burnouts or other types of similar difficulties, it's probably pointless.

I used many cards (normal front/back way) with more than 95% results on ANKI but never really being able to use them in context or seeing any valuable result.

Then I learnt that probably simply drilling decks wasn't the right strategy to use to get real results. And if you don't tailor them in the proper way you could have the illusion to learn something when in reality you won't. Same things with app like Duolingo or similar.

Compared to LingQ were if you are constant and keep going you will see effective results. The mind will have many more associations and inputs to increase its own ability to progress.

If you use Anki for L1->L2 retrieval operations, it improves the use of the L2 (your so called "active vocabulary") as well. And that's not always (at least "directly") the case with reading and / or listening alone because the recognition operations only tend to act as a mechanism to facilitate the use of language...

Regarding this, could we say that this retrieval operation training is more effective with cloze cards? And also tailored to what your mind really need to boost its own capacity to improve the language?

Why not skip L1->L2 retrieval operations using ASRS and resort directly to speaking and writing activities = language use?

Yes, that's "the" question in the age of generative AIs à la ChatGPT and Co that seem to get better by the minute.

Yes, I believe this is a better strategy and also gives you a more complete use of the language. I can see writing improve quite a lot the capacity of the mind to recall many things (and in context and in real own usage) but this also requires more mental energy. When you have more vocabulary it becomes easier to manage.

I believe there is still a place à la ANKI, it's just not so random. The more precise and structured the creation of the deck and the type of recall, the better results you'll have.

But is this really necessary? I believe it's not unless the student has specific needs that we saw already in other answers.

The fact is that a generative AI (inside LingQ) could not only improve the experience for everything we said already but could also intervene in the creation of the right deck to study based on our own vocabulary and needs. It's just statistics and it will save a lot of time, plus it'll give the right phrases and words to drill. It would be a massive time improvement.

xxdb

gb   United Kingdom
May 23 at 13:13

TLDR response to your post: my personal gut feel is that anki vocab in combo with material which includes those words will cement the words faster and with less effort into your long term memory.

Longer version:

I've done the experiment with a hard language (as some of you know).

What definitely *won't* work is using anki with specific vocabulary which does not match what you are reading. Essentially it's two different tracks with no complementarity. Learning a vocabulary list in anki and reading random material in lingQ doesn't overlap enough to get any kind of synergy. I know that because I tried it.

What does seem to work to give you synergy is to extract your daily anki words and plug them in to your daily lingQ reading by using AI to generate stories including those words.

Stats wise: doing it the less efficient way I was about 50-60% retention rate on anki (sometimes worse - below breakeven) for a hard language (russian) while I was just reading children's novels in lingQ. Sitting at around breakeven point means you would be "churning" as in the words don't get fully burned into your long term memory - they drop out and you have to do them again. Which obviously defeats the purpose of anki which is to *memorize* the material.

This is compared with 80-90% retention rate with anki in French while reading exactly the same novels in French.

Now... my stats have gone up to 65-75% fluctuating but on a slow trend upwards of about .1% per day since I've been using chatGPT to generate vocabulary specific stories for me on a daily basis.

This means that there is now the potential that I can get to "done" in a reasonable time frame (i.e. small number of months instead of years).

In this case "done" means enough words "memorized" that I can consider the memorization phase of the language acquisition process to be over with and I can just concentrate on using the language.

nfera

au   Australia
May 23 at 15:30

@xxdb

"What definitely *won't* work is using anki with specific vocabulary which does not match what you are reading. Essentially it's two different tracks with no complementarity."

I mentioned it on your other thread, but I think this could've been because your LingQ reading material was children's fantasy novels (like Treasure Island). If your LingQ material had of been conversations or graded readers (at the appropriate at which you were up to on your frequency list in Anki, eg. word idk 5,000), you would've had a different experience. The graded readers and your Anki cards would be using a very similar frequency list, so you would've been seeing similar words. Especially, if you were using LingQ for extensive reading, instead of intensive reading, as you would've been seeing much more content.

But, as your experience highlights, it's definitely possible to do two separate things, which are out of sync and don't complement each other much. That is, drilling high- and -mid-frequency words on Anki while doing intensive reading on fiction books (i.e. with many low-frequency words).

Obviously, to ensure the two activities are synced, you can do what you are doing, pasting your flashcard words in ChatGPT. The other way is to use the LingQ -> Anki option. Then the two activities would be in sync. Though, if you use graded readers at the appropriate level corresponding to your Anki frequency list (with native audio, or TTS as you are doing with your ChatGPT stories), they would also be in sync, and most likely more interesting content and corrected by native speakers instead of a bot.

nfera

au   Australia
May 23 at 11:25

@PeterBormann

I do wonder to what extent shadowing would provide a similar effect. If your goal of Anki is L1/NL -> L2/TL in order to practise using the language/vocabulary, so you can later produce it, shadowing to some extent may provide a similar benefit.

One issue which I've noticed in Italian is that I can recognise a word pretty much 100% of the time, but not necessary say it. I suspect you experience a similar thing. I can say half the word, or I partially can produce the word, but maybe I mix up the ending or the second half. Eg. If the word is, say, 'politica', I might try and speak it and say 'politisa' or 'politicono' or some nonsensical word instead, but half-way there. The word is 100% recognised, but only idk 50% able to be produced. This is because I don't know the word well enough and have most likely never spoken the word before. Often,, after I've said the word one or two or five times in a conversation, I won't mispronounce it/say the wrong word again.

As production reinforces your 'knowledge'/familiarity of a word/the vocabulary, I'm wondering to what degree shadowing would provide. The obvious guess would be not as much as mentally-straining production, but better than nothing. The question is to what degree in there. Would it be an efficient way to go from these 100%-recognised-but-only-50%-able-to-be-produced words to 100%-known-and-able-to-be-produced? The added benefit is obviously pronunciation practice.

xxdb

gb   United Kingdom
May 23 at 13:15

Glossika is built around this concept. I do believe that glossika "works" in that you end up knowing all the phrases. Allegedly this gives you a "model" to work with so that you can build other sentences with these models as a base line. I can see logically how it would work but I've never followed through and my subscription expired so I can't say for sure.

davideroccato

it   Italy
May 23 at 16:34

The word is 100% recognised, but only idk 50% able to be produced. This is because I don't know the word well enough and have most likely never spoken the word before. Often,, after I've said the word one or two or five times in a conversation, I won't mispronounce it/say the wrong word again.

I give you an answer but I'm not really sure exactly about what the problem here is.

The fact is that I don't understand if you recognise the word by reading but don't remember it when you use it.

Or if you don't remember the use of the word in context but it seems you do.

Or if you don't recall the spelling correctly and this is why you mispronounce it.

So my answer could be completely wrong. I apologise for that.

I'm not a fan of shadowing at all but I don't think this is the problem here. If you recognise the words 100%, do you know how to pronounce them correctly first? And I mean also in your inner vocalization when you read daily?

In Italian language, like German for example, pronunciation is not so difficult per se. You pronounce the words as you see them. A part from the accents that written italian language doesn't use much.

If you learn the alphabet very well, you cannot mispronounce "politica" IF you recognise the word in your mind. Even if you never used it before but you have it in your mind.

If you recall the spelling of the word when you talk you should learn the pronunciation of the alphabet very well and artiiiiiicuuuuulaaaaate a bit. Try reading by articulating very slowly and stretching the words very slowly. (I did a lot with French and helped a lot).

You can't be wrong on ca-sa, po-li-ti-ca, gra-fi-ca, pa-let-ta, ri-ga, etc.

You might have difficulties on not being sure about po-lí-ti-ca or po-li-tí-ca etc. but not on other things. I suppose the last one will come with time as soon as you will get the "sound" of the language.

If you don't recall the spelling, that's something else but shadowing won't help either. Imho.

nfera

au   Australia
May 24 at 11:03

@davideroccato

I'm not referring to an issue with the spelling, but I'm referring to the example @Anxxos gave, where you can recognise the word, but not produce it correctly. I don't do any subvocalisation, as you seem to do (this is because I read while listening), so this may the difference.

I find that if I (fake example here) read while listen to a word 10x, I can recognise the word and know the definition. I wouldn't be able to produce the word though. I may know the word starts with, say, a 'b', but I won't be able to tell you the word. But If I read while listening to that word another 40x, I'll be able to produce it. So there's this discrepancy between 10x read while listens = I can recognise the word, but it requires an additional 40x read while listens (i.e. 50x in total) to be able to produce the word.

My thinking was to of how to shortcut the required number of read while listens to be able to produce the word. I find if I can already recognise the word, then I say it out loud idk on 5 separate occassions, I can then independently produce the word. So I'm pondering if a way to shortcut this, to be able to produce it, is 10x read while listens to be able to recognise the word, but then shadowing to reach those 5x required to be able to produce the word. And this would be faster than 10x read while listens PLUS 40x read while listens. The goal here is how to reach a state where you can produce the words. The secondary effect of shadowing is obviously increasing one's poor pronunciation.

davideroccato

it   Italy
May 24 at 11:43

@nfera: I read @Anxxos or @Peter but I probably don't understand it as a common thing because it seems I don't have this specific problematic (compared to remembering grammar rules for example!). Or my brain seems to use another strategy for this particular effect that I workaround it. I never thought about it.

This is connected to what we were talking about the other thread about visualization and not many seemed to give it to much importance or analysis. And writing connected to visualization.

What happens to me is that if I talk and recall the word I use it. If I don't recall it I don't use it.

If for any reason I might get stuck because I don't remember a word my brain switches to another possible word. Which is a strategy that seems to be very common amongst many polyglots. The capacity to quickly choose words that we can workaround in different situation avoiding embarrassing cul-de-sac.

If the problem is not spelling than I see it as a visualization problem and I'm afraid I don't see ANKI as a solution. And of course, I might be wrong but you do already lots of repetitions and doing more of the same thing it doesn't seem to be the right answer to me.

In both examples, if you know "politica" or "pasticceria" you can't be wrong. If you see the word in your mind. So I see it as a difficulty to visualize in your mind, when you speak, the words that you already know.

Try to observe this mechanism for the next times you have it. Do you see the word in your mind? Does it take too long to appear? Do you see just a part of the word?

In order to reduce the number of repetitions you are doing I would try to reinforce your capacity to recall with visualization. It's always a recall and of course you might use ANKI for that as well but not with languages but with images or other things. It would be a training for your mind that you can use for anything else too.

I suggested you about writing the other time because with writing you SLOW down the process allowing to your eyes to absorb the letters.

You can also slow down while reading+listening and instead of focusing only on increasing vocabulary or improving listening, you can pay more attention to visualizing the letters of the words.

For example, sometimes I listen to 0.9x even if I can go 1x or 1.25x but I pay attention to the words or other grammar structures while reading and listening. So my eyes will have more time to process more information.

It was very easy to do it in the past, unfortunately I don't have the energy to study much in these last years otherwise I would do a lot more.

I hope I gave you some other food for your brain to consider. In any case, if it's not the right solution, it doesn't hurt. :D

Learning is a never ending experience.

nfera

au   Australia
May 24 at 12:11

@davideroccato

It seems you don't have the experience, as you say. Think of it like this: If I ask you to name all the members of Axis in WWII. You may remember Nazi Germany, Italy, Japan, Bulgaria, Hungary, Romania, Slovakia, etc., but you forgot that Yugoslavia was a member of Axis for two days. This is the situation. I remember most of the word, but don't remember a small part of it. I couldn't tell you the spelling nor how to pronounce it. Like the list of countries in Axis, I can only produce most of the list of sounds/letters in a word, but not the entire list of sounds/letters.

PeterBormann

de   Germany
May 23 at 20:21

Hi nfera,

Good question!

I will answer at the WE. Unfortunately I'm very busy at the moment and don't have time to participate in discussions in the LingQ forum...

Eine gute Zeit & bis dahin

~Peter

PS -

My main problem is "Dutch" right now.

Recognizing (esp. with a lot of guessing) is super-easy (going from A0 -> B1 in reading / listening comprehension was just a matter of a few hours), but producing / using the correct Dutch words / phrases is next to impossible :-)

Anxxos

je   Jersey
May 24 at 10:06

Yes, @nfera raises a crucial point about the cognition of a foreign language:

It's not enough to simply know how to translate 'X' from a foreign language; your brain must also be able to recall 'X' when you need to express yourself in said foreign language.

I can provide a personal example. I never had to look up the meaning of 'panetteria' to understand it. 'Pane' equals bread, '-eria' is equivalent to '-ery', thus 'bakery'. However, the first time I tried to recall it while speaking Italian, I said 'panecceria'... a silly and typically inaccurate recall. My brain had memorized the beginning and the end of the Italian word, but had forgotten the not-semantically-useful-to-me-when-decoding-from-Italian middle part.

Our neurons are hard-wired: learning a foreign language requires bidirectional activation. Yes, you must imprint it biologically twice. Learn L2>L1 and learn L1>L2 (at least if you want to speak and write L2). Lingq is only great at the former.

An important corollary is that your learning speed might prove massively different between the two directions, as PeterBormann rightly highlights. Requiring a very different level of effort AND PAIN.

One major drawback of Lingq's SRS offering is that it presents the two flashcards L1>L2 and L2>L1 on a given word on the same day! Oh no! This is a mistake! It's vital to separate the learning of the two.

Thankfully, Anki does this by design, and it will automatically make you work harder on the more challenging aspect: translating from native into foreign.

nfera

au   Australia
May 24 at 11:07

@Anxxos You hit the nail on the head.

I think bidirectional activation is not 100% needed per se. For instance, if you hear the word 'ciao' a million times, for instance, then you can just say it the first time. But, as per my above comment, I think it requires a disporportional amount of exposures to go from recognition to production without practising production (in the example 10x for recognition, then a further 40x for production). This is where practising production comes in to be a faster way to be able to independently produce the word.

hiptothehop

us   United States
May 25 at 03:45

@ (Sorry, I can't reply directly) Taking your "panetteria" example, do you expect it is worth the effort to learn/activate such words learn L1>L2 with repetition on Anki? Or, is the experience of not being able to produce it correctly once, and then learning it at that time not sufficient, and perhaps more efficient?

I may be too comfortable with my own ignorance or making mistakes :-), but it seems like having someone correct me or ask them to remind me of the word for "bread store" in the L2 at exactly the moment I need it, could be more efficient than spend time drilling with Anki L1>L2?

Thanks for your posts.

nfera

au   Australia
May 25 at 10:09

@hiptothehop

"Or, is the experience of not being able to produce it correctly once, and then learning it at that time not sufficient, and perhaps more efficient?

I may be too comfortable with my own ignorance or making mistakes :-)"

Indeed, it probably is a more emotional situation, so the memory will burn much stronger and therefore be a faster way. The issue that I have is that it's with maaaany words and so sometimes what I say isn't very comprehensible, when I'm speaking to people. :P Furthermore, with regard to @Anxxos' example, you may remember the first third and the third third, but forget the middle third of the word. But this happens in a wide range of degrees. For instance, you may only remember the first letter/sound of the word and nothing else for other words. One word here and there is no problem, as in the conversation, you can just ask for the word that means 'bread shop' is, but with many words, it just becomes a real issue. This is usually on topics, you've never talked about before. Eg. Your first few times you talk about philosophy or politics, etc. It obviously gets better over time, but we are looking for a way to speed up the process.

PeterBormann

de   Germany
May 21 at 14:04

"With no special prior reading practice."

Hm, according to your LingQ stats (hello, Michilini! No, we won't have "that" discussion again and again and ... :-) ), you've read more than 6.6 million words so far. I think you mentioned that you did the language tests in 2022 (?), so you had "extensive" reading practice prior to those exams :-)

You'll also likely pass a C1 level reading comprehension exam - without any special preparation...

Reminds me of a saying in German:

Es ist nicht nötig, Dein Licht unter einen Scheffel zu stellen :-)

https://www.geo.de/geolino/redewendungen/6023-rtkl-redewendung-sein-licht-unter-den-scheffel-stellen

Anxxos

je   Jersey
May 23 at 10:17

My two cents, after learning and experimenting with foreign language acquisition through SRS software like Anki and Supermemo for over 20 years, is as follows:

The fastest progress is ensured by rote learning of simple sentences in the target language, ranked by word frequency.

Be ruthless. Give yourself five seconds to produce the sentence in the foreign language when you see the flashcard's sentence in your native language. If it doesn't quickly come to mind, you don't know it well enough. Choose 'again' or 'fail' and start over until it flows naturally.

Pain and gain.

davideroccato

it   Italy
May 23 at 10:32

Pain and gain!

image

The fastest progress is ensured by rote learning of simple sentences in the target language, ranked by word frequency.

Part of this can indeed be optimized with LingQ by reading common material + tagging phrases + exporting to ANKI (if necessary). If only LingQ would increase the limit of 9 words!!!

xxdb

gb   United Kingdom
May 23 at 13:02

I think you have to use the language for your active vocabulary to kick in regardless of whether you use anki or not. I can't say in my own case because I stopped using anki in Spanish when I hit the point I thought I had enough words and I immersed. I haven't immersed in either French or Russian. My spoken Russian sucks. Like it's *bad*. But my French was functional even though all I did was anki, youtube and lingQ. I was in a French speaking place last week and I managed to have a decent conversation with a taxi driver about a bunch of things. I would probably not pass any kind of exam though.

Other anecdotal evidence of anki's use is that English dude (see the relevant michilini post) who immersed but who used anki for words/phrases he ran into during conversations.

I dunno.

I'm inclined to think that if you are well into learning a language without anki you can probably survive without it.

On the other hand if you are starting out, well I believe it works for me and there are others who say the same thing, so I would say why not in that case.

David72

gb   United Kingdom
May 21 at 08:15

Hi JBelly. I used Anki for a couple of years and gave it up 2 months ago. Best decision I've made!

I found I am learning more by using my time to read and listen to real content in context. Plus, it's more fun. Anki was a bit of a chore - like having a school test every day. I am absorbing and retaining words, constructions, phrases, grammar much better by seeing them in context in many different types of dialogue and scenarios than memorising them in isolation.

davideroccato

it   Italy
May 20 at 14:22

@JBelly: ok, maybe I'm gonna give you the voice you were waiting for. :D - Based on my opinion only. ;)

If you like LingQ and want to simplify your life by avoiding using too many tools that are not really necessary to your goals, just do that!

The fact that many people talk about something doesn't mean it's useful for everyone.

Plus we have a lot more tools today than when ANKI was initially created.

Others have already said why ANKI was created.

Repetition is important but with LingQ you do that naturally. And the most common words will come up many different times anyway so you'll keep reinforce them.

You can use LingQ even more strategically by exploiting repetitions in many different ways. You can repeat books, lessons, playlists...

The question is: "Why should you use or add ANKI?"

If you use LingQ better and better you shouldn't in most cases. It's a matter of personal preferences.

I used to use it for a long time and I don't anymore. Mostly because you need to use it strategically, not every repetition using ANKI work, I have drilled cards for a long time and didn't get any benefit because I didn't do them properly. (I figured that out afterwards)

Using ANKI very well would require time to tailor it to your needs. And a learning curve to adjust it to yourself.

The more time you dedicate to it and more precise are your cards, the better results. For example, creating cloze cards would be one of most reliable methods to stimulate your memory recalls.

Plus you need some extra mental energy for that specific exercise as well.

Some example to consider:

For example, you want to accelerate the first 1000 common words at the beginning of a language. (It's not necessary, you'll get there anyway with patience but for some language could be a nice boost).

You want to quickly build a vocabulary for a specific professional field because you don't have much time available.

You want to improve (for some reason) your target language words that you don't see often. I would wait to use it after you reach a high number of words in your target language using LingQ and then see if it's really necessary. For example, for Arabic, I would wait until 70k known words with 95% contextual comprehension. (that's my opinion!)

(Most interesting to me at the moment)

You would want accelerate or improve collocations, some grammar structure, some other words combinations that are important and that for some reason you don't see often or have a hard time to digest.

(I would consider that after the same higher number of words mentioned above, after having listened and repeated many hours of content, after writing and speaking many hours as well.)

LingQ's SRS system is not bad but the inner vocabulary structure could be very useful if used correctly.

You could start tagging words and expressions or phrases that you find interesting in LingQ and build your own future reference. You could eventually use them for ANKI in the future.

Plus today other users are experimenting this with ChatGPT and creating stories with these words, and I would prefer this method.

ANKI is a wonderful powerful tool (as @Peter and @Anxxos said) but for a serious LingQ user become less interesting compared to people that don't use LingQ (or a similar method) or use it only randomly.

If I had to think about a companion to LingQ today, I would rather learn how to use ChatGPT first for all those words or situations above and import the results to LingQ. Then I would think about ANKI for what's left if it's still necessary for your OWN goals.

IMHO.

PeterBormann

de   Germany
May 20 at 14:40

"ok, maybe I'm gonna give you the voice you were waiting for."

Davide, you mean the "voice of wisdom we have all been waiting for"? :-)

Anki or ChatGPT?

"If I had to think about a companion to LingQ today, I would rather learn how to use ChatGPT first"

Yessssssssssssss...

Or in my case the combo:

LingQ + generative AI à la ChatGPT (used with promising prompts for SLA) + Anki (via LingQ-to-Anki-exports)

A nice WE to you all,

Peter out

davideroccato

it   Italy
May 20 at 15:52

Naaa, just the voice she (her inner voice) was secretly waiting for inside her heart that was saying/asking: "Please, please, I'm a happy LingQ user, I really really like what I do, please help me, tell me I don't need ANKI for now and I can keep learning with LingQ". :D

That voice.

However, after reading some fundamental scientific research, I added a :D to reinforce my opinion. Making me definitely much wiser.

generative AI à la ChatGPT

I like this expression, it's more accurate, I often mention ChatGPT but what I would like to convey is exactly that.

JBelly

gb   United Kingdom
May 20 at 12:06

Thanks for advice/suggestions. No issues with time (retired) or enthusiasm/dedication (life/family keep me from learning/reading all day, every day) but I’m still at a loss as to why Anki rather than just LingQ’s SRS system - is there really any difference?

PeterBormann

de   Germany
May 20 at 13:33

Anxxos is right (see his comment below):

LingQ's SRS is rudimentary compared to Anki and not that useful.

It probably shouldn't even exist because Steve doesn't recommend its use

(but, of course, an ASRS isn't compatible with his "fun first" learning philosophy,

which is problematic in itself - at least from my teaching perspective).

Be that as it may, most LingQ users probably don't use the SRS and stick to the main functions of LingQ, namely: reading and/or listening. So they wouldn't notice if LingQ's SRS suddenly disappeared.

Anki, on the other hand, is extremely powerful - to the point of being overwhelming.

Therefore, it should be used with care...

PS -

Here's a basic Anki user guide aimed at beginners based on this reddit thread (https://www.reddit.com/r/Anki/comments/ycnnb2/basic_user_guide_for_anki/):

https://docs.google.com/document/d/1HiO1Fm3RLOmZiL0tsTXndY4TmzMiZFMQeENx5uCXVE0/edit

PPS -

There are "shared Anki decks", for ex. for Arabic, as well:

https://duckduckgo.com/?t=ffab&q=anki+decks+arabic&atb=v343-1&ia=web

But I've no idea how good the quality is...

However, you could also buy pre-made Anki decks for Arabic from lingualism.com:

https://lingualism.com/

PeterBormann

de   Germany
May 20 at 10:42

BTW, here is a nice article about the topic by one of my favorite polyglots, Luca Lampariello: https://www.lucalampariello.com/why-i-dont-use-anki/

But his "no's or con's" are not valid when LingQ's "reading (or / while) listening" approach is coupled with the "LingQ-to-Anki" function:

"Here’s five reasons why" (PB: we shouldn't use Anki):

"Making flashcards wastes learning time"

No. LingQ does it for me.

"Adding new cards can become an addiction"

No. LingQ does it for me. Besides, there's a difference

between "habit" and "addiction", but that's another story :-)

"Reviewing old cards can become a chore"

Just reduce the number of flashcards to be reviewed.

"Flashcards take language out of context"

Yes, "acontextual / no sentence" flashcards are as bad as

simple word lists "Katze (in German) - cat, "saß" - sat, etc.".

However, LingQ provides context clues as well...

"Brain-friendly learning strategies make SRS apps irrelevant"

No. Just use the SRS as complements, for ex. combined with

reading / listening / speaking / writing activities.

I like Luca, but there's a general lesson to be learned here:

never trust anyone blindly (not even yourself or your fav AI) :-)

Anxxos

je   Jersey
May 20 at 15:38

I could not agree more: that silly, anti-SRS hit job was rather surprising and disappointing from a clever guy like Luca.

But, hey, the guy wants you to BUY his language method...

PeterBormann

de   Germany
May 20 at 10:04

Hi ,

It's not about "Anki" per se, but about the usefulness of "Artificial Spaced Repetition Systems (ASRS) " in general compared to reading as a kind of "Natural Spaced Repetition System".

Yes, ASRS definitely work. See, for example: https://en.wikipedia.org/wiki/Spaced_repetition

And they can be used at all language levels by:

  • beginners: starting with the most frequent vocabulary to get a boost (see, for ex., the free study decks on Memrise).
  • intermediate learners: in my case, for testing purposes (i.e., in L1 -> L2 retrieval operations, which are more difficult than L2->L1 recognition operations) using LingQ-to-Anki exports.
  • advanced learners who try to memorize less frequent, but conventionalized word groups (collocations).

The main advantage of "reading (and esp. reading while) or listening" is: it's more enjoyable (esp. from a low intermediate language level upwards) and context-dense than ASRS.

However, ASRS can be quite effective and efficient - provided that enough context information is given -, but it takes some time getting used to them and requires a higher frustration level.

So, to make a longer story short:

It depends on your

  • learning style,
  • frustration level,
  • expectations,
  • study routine,
  • and time budget.

As a rule, it's not a good idea to completely replace "reading (or / while) listening" with SRS, but the latter can complement the former quite nicely (although SRS activities should be limited to 5-10 minutes per day).

If I were you, I'd do a little experiment (say 14 days) with the LingQ-to-Anki export function using Anki (just download the LingQ flashcards and import them into Anki). Then you can get a feel for how the whole process works and how you can integrate it into your usual study routine.

Hope that helps,

Peter

Anxxos

je   Jersey
May 20 at 07:37

There is little scientifically controlled research behind Spaced Repetition Systems (SRS).

One of the few scientists who have presented data is Piotr Woźniak, the founder of SuperMemo. Anki uses one of its algorithms, known as SM-2.

The algorithm used by LingQ appears to be more rudimentary, particularly because it doesn't adapt the repetition intervals based on the difficulty you experience in learning a specific card. It also doesn't offer long-term repetitions, which is detrimental to long-term recall.

As a result, for the same amount of effort (the same number of flashcards reviewed), you're likely to memorize less. The gap is probably significant.

Given that SM-2 is freely available, it's regrettable that LingQ hasn't implemented it. This would be the logical step to benefit its users in my opinion, but it perhaps reflects the lack of interest in SRS by its founder, Steve Kaufman.

If you start using Anki, you'll quickly feel the adaptive nature of the algorithm. Your decision to stick with it will likely depend on your confidence that the extra effort (exporting to Anki) is worth the enhanced memorization.

Based on my experience using SRS since 1997, adoption depends on personal preference and the acceptance of new, disciplined cognitive habits.

Not everyone will enjoy it, and Steve Kaufman's more relaxed approach (emphasizing massive exposure to comprehensible input instead of using flashcards) may become a reasonable compromise.

I personally use Anki every day of my life. It's always open on my desktop, and I continually add to it anything I want to remember. A life companion!

The phone app is excellent.

PeterBormann

de   Germany
May 20 at 09:28

Hm, Anxxos, not sure what "little" means in this context.

That the "spacing effect" is beneficial to learning has been studied for almost 150 years (since Ebbinghaus 1885): https://en.wikipedia.org/wiki/Spacing_effect

And spacing is usually seen as the superior recall technique compared to cramming.

Besides, if you use "spaced repetition" as search terms on Google Scholar, you'll get more than 5600 search results: https://scholar.google.com/scholar?hl=de&as_sdt=0%2C5&q=%22spaced+repetition+%22&btnG=

Yes, that's not drowning in search results, but "little" looks a little different... :-)

PS -

"A life companion!"

Same here :-)

xxdb

gb   United Kingdom
May 23 at 12:36

My take on it is that it provides much more reliable "proof" that a word is actually "learned". In LingQ it's highly subjective.

In anki, if your repetition gap is more than one year for most of your 10,000 most frequent words, it is as close to a scientific measurement as you're going to get that you have memorized those 10,000 words.

xxdb

gb   United Kingdom
May 23 at 13:26

Yeah. The app I used for Spanish was SM, not anki.

Page 1 of 1

WANT TO LEARN A NEW LANGUAGE?

Learn from content you love!