Here's how many words and how much time you need (probably)

General discussion about learning languages
User avatar
jeff_lindqvist
Black Belt - 3rd Dan
Posts: 3135
Joined: Sun Aug 16, 2015 9:52 pm
Languages: sv, en
de, es
ga, eo
---
fi, yue, ro, tp, cy, kw, pt, sk
Language Log: viewtopic.php?f=15&t=2773
x 10462

Re: Here's how many words and how much time you need (probably)

Postby jeff_lindqvist » Mon Jul 26, 2021 4:58 pm

luke wrote:There is also the assumption that someone can learn 500 words a month and keep that pace going for almost 3 years. Those with Anki vocab experience know this projection may be optimistic and unrealistic.


It would be an interesting challenge, though. ;) 500 words per month means ~16.66 words per day. Perhaps there is a decent deck out there, with exactly 16000 words... Assuming they are all introduced at the right time, learned equally easily (to the level where you'd select "Easy" on the second run-through on the first day you see them, and every consecutive time you see them), you use the default interval settings, and most importantly - never miss a single day - how many words would you have to review on the final day? (It's probably dead easy to calculate this, but I'm too tired at the moment...)
2 x
Leabhair/Greannáin léite as Gaeilge: 9 / 18
Ar an seastán oíche: Oileán an Órchiste
Duolingo - finished trees: sp/ga/de/fr/pt/it
Finnish with extra pain : 100 / 100

Llorg Blog - Wiki - Discord

gsbod
Blue Belt
Posts: 839
Joined: Sat Jul 18, 2015 3:22 pm
Location: UK
Languages: English (native)
German (advanced)
French (intermediate)
Japanese (intermediate)
Spanish (learning)
Language Log: viewtopic.php?t=1152
x 2900

Re: Here's how many words and how much time you need (probably)

Postby gsbod » Mon Jul 26, 2021 5:15 pm

It's not precise, but as a rule of thumb I find that my Anki reviews are around 10x the number of new cards introduced a day. So if we're assuming 16.66... new words a day translates to 33.33... new cards a day, you'll be looking at a typical workload of 333 reviews + 33 new cards. Which is too much for me, at least!

Not that I would ever expect to continuously use Anki, with the same deck, from A0 through to C2.

Edit: Sorry, it's 10x new words added a day, not cards, assuming that each word is learned with two cards (so you practice in both directions). So we're talking 166 reviews a day plus 33 new cards, which is still too much for me long term (although I have done it successfully over a short period of a few weeks before).
2 x

Beli Tsar
Green Belt
Posts: 384
Joined: Mon Oct 22, 2018 3:59 pm
Languages: English (N), Ancient Greek (intermediate reading), Latin (Beginner) Farsi (Beginner), Biblical Hebrew (Beginner)
Language Log: https://forum.language-learners.org/vie ... =15&t=9548
x 1294

Re: Here's how many words and how much time you need (probably)

Postby Beli Tsar » Tue Jul 27, 2021 11:31 am

jeff_lindqvist wrote:
luke wrote:There is also the assumption that someone can learn 500 words a month and keep that pace going for almost 3 years. Those with Anki vocab experience know this projection may be optimistic and unrealistic.


It would be an interesting challenge, though. ;) 500 words per month means ~16.66 words per day. Perhaps there is a decent deck out there, with exactly 16000 words... Assuming they are all introduced at the right time, learned equally easily (to the level where you'd select "Easy" on the second run-through on the first day you see them, and every consecutive time you see them), you use the default interval settings, and most importantly - never miss a single day - how many words would you have to review on the final day? (It's probably dead easy to calculate this, but I'm too tired at the moment...)

I suspect it's not quite as easy to simulate as it seems... But running it through the (somewhat unpredictable!) Anki simulator, I get the following figures. I've based it on the settings and retention rates for the 2000-odd Latin words I've learned in the last few months, and rounded it to 1000 days. Note that this isn't modelled with the simplifications you suggest above - it's modelled more accurately using my own learning steps, interval modifier, and retention rate at every step, as it's been for those 2000 words.
  1. 16 cards a day, recognition only:
    - Maximum reps 513 (not on the final day... Anki has more randomness in it than that!)
    - Average reps 363
  1. 32 cards a day, both directions:
    - Maximum reps 975
    - average reps 723

Interestingly, the recognition-only average reps is slightly less than my own average reps for the last three months, which has not been a period of intensive learning. I don't think that this level is particularly unsustainable... Using averages worked out from my own word cards, this would take about 15 minutes a day, which isn't bad at all. The issue would be also maintaining enough input and output to make sure those words stuck in a usable way, especially the later, rarer words.

The two-direction one is harder to model, without a deck that does this directly. Some of the cards would be a lot easier, because you get them twice, and some are harder because they are active. Would it also get disproportionately harder as you go on, since later infrequent words are less easy to turn into active vocab?

Realistically, even with one direction, your retention rate is going to be very different at different stages - low for the first thousand, then increasing rapidly as you get used to the language, and then at some point hitting much rarer words and so dropping again. I've less experience of this far end of the bell curve, but have hit it in Greek, so that individual word cards are much less effective than at earlier levels. If I read more and more widely this would be less of an issue, of course.

The other issue is that if you were enough of an Anki addict to keep going at that level for 1000 days, you are probably not the kind of person who can resist sneaking in some extra Anki drills - grammar, sentences, whatever - that would really slow you down. Which is of course exactly what I have done and why I've averaged 10 words a day for the last 6 months rather than 16.

All this presumes a reasonably familiar language. I don't have the stats directly for my own Farsi experience, but it took me, as I recall, basically double the time to learn an equivalent number of words. Latin might not be a category 1/2 language, if it had a category, but the vocabulary is as easy to an English speaker as a romance language. Just Anki-ing your way through a category 4 or 5 language would be seriously tough. There's a reason all the AJATT learners of Japanese use vast volumes of sentence cards - the words are just much harder to pick up in isolation than a related language would be.
4 x
: 0 / 50 1/2 Super Challenge - Latin Reading
: 0 / 50 1/2 Super Challenge - Latin 'Films'

jeffers
Blue Belt
Posts: 848
Joined: Sat Aug 22, 2015 4:12 pm
Location: UK
Languages: Speaks: English (N), Hindi (A2-B1)

Learning: The above, plus French (A2-B1), German (A1), Ancient Greek (?), Sanskrit (beginner)
Language Log: https://forum.language-learners.org/vie ... 15&t=19785
x 2774
Contact:

Re: Here's how many words and how much time you need (probably)

Postby jeffers » Tue Jul 27, 2021 1:04 pm

GandwolfTheGrey wrote:Time needed per level:

A1 - 1 month
A2 - 2 months
B1 - 4 months
B2 - 8 months
C1 - 16 months
C2 - 32 months


Time and again, new members have joined the forum with very high goals, such as reaching B2 from nothing within a year, and so on. They start a log, explain that for a couple of weeks they have been studying 2/3/6 hours per day and see no reason why they couldn't continue at this pace. They get a lot of encouragement, as well as the usual advice about courses and making sure they include native content. Whenever one of these logs is started in one of my languages of interest I subscribe to the topic, to see how it goes. Unfortunately, I have yet to see one of these logs carry on to the goal. It's possible that some of these users met their goal and just didn't post on their log, but I suspect that they all burnt out within a month or two.

My point is that we are flesh and blood; calculations of what can be done efficiently are fine for machines but rarely take into account the human factor. A lot of "time needed" calculations are based on courses like FSI, but I think there is a big difference between being on an intensive course and studying in your own time. On a course like FSI you have a timetable of lessons and homework; in addition you have other people who you see every day, students and teachers, who are helping you and whose expectations you don't want to disappoint. I do believe that FSI know what they are saying when they say that their students can reach X level in Y time, but I have yet to see a self learner achieve the same results.

I do find the vocabulary targets useful. Having a general idea of the vocabulary size needed for my target level is useful in the same way as knowing what grammar I will be expected to use and understand, knowing how much reading, writing, listening and speaking I would be expected to do and what types of things I would be expected to read, write about, hear about or speak about. Just cramming the right amount of vocabulary isn't going to gain you the level, but it is one small part of knowing how you are progressing.

I think people considering learning a language need to know that they are taking on a massive task. Telling them that they could get from 0 to B2 in 8 months is not only wrong (for the vast majority of people anyway), but it sets up unrealistic expectations. In fact, if a person wants to learn a language well, they need to put in the time, but also take their time because it is a long road.

Finally, the quicker something is learned the quicker it can be forgotten, or so I've heard. Take your time and you will reap more benefits than simply achieving the goal, you will have a deeper and more lasting understanding of what you have learned along the way. Slow learning is deep learning (which I think was the title of one of my old logs :lol: ).
7 x
Le mieux est l'ennemi du bien (roughly, the perfect is the enemy of the good)

French SC Books: 0 / 5000 (0/5000 pp)
French SC Films: 0 / 9000 (0/9000 mins)

User avatar
Le Baron
Black Belt - 3rd Dan
Posts: 3510
Joined: Mon Jan 18, 2021 5:14 pm
Location: Koude kikkerland
Languages: English (N), fr, nl, de, eo, Sranantongo,
Maintaining: es, swahili.
Language Log: https://forum.language-learners.org/vie ... 15&t=18796
x 9389

Re: Here's how many words and how much time you need (probably)

Postby Le Baron » Tue Jul 27, 2021 3:38 pm

^ I agree wholeheartedly with the above. Especially about the constant failure to recognise the 'human factor' when plotting graphs and running algorithms, which reeks of the delusional idea that solving technical problems means it's plain sailing for human practise. The hard fact is that you can plan a day out with what you think is military precision and it might still rain, plus you have no control over things you don't yet know and thus can't asses. Causes unnecessary frustration.

Which means a long-term goal is probably best left in outline (and essentially put to the back of your mind and humoured) whilst you get on with the actual steps of that journey. Tiresome obsession with mathematical modelling is something that has poisoned my former profession of economics, as people see ideal worlds in models (also informed by erroneous conceptions of real-word operations anyway) which then fail to tally with practise.
But back to language learning... I think Jeffers is right when pointing to people who are more savouring the idea of being 'B2' before the actual journey has even really begun. It's not that it's bad to visualise success, but factoring the amount of time and work, plus setbacks, has to be realistic. Those crazy projections of learning 500 words a month ignores reality; that when you divide up your words to 16 or so a day you won't have mastered some (many?) of those within one week and then you're already 'behind'. There's surely a difference between motivation and unrealistic pressure? And as was also alluded to, why do the people who propound these methodologies hardly ever have real all-round proficiency in multiple languages? It's a damning clue.

I rather admire people like Carmody and others here who are largely dedicated to one language. I'll freely admit that I'm average and there are people cleverer and more able than me who can learn more languages more quickly than me, but this is not the norm and there's no mathematical model to follow which can be used for the average student to follow which will deliver it. It's a polyglot affliction really: desire for multiple language proficiency + impatience + awareness of limited time = search for the magic chalice of superfast learning.
6 x

Beli Tsar
Green Belt
Posts: 384
Joined: Mon Oct 22, 2018 3:59 pm
Languages: English (N), Ancient Greek (intermediate reading), Latin (Beginner) Farsi (Beginner), Biblical Hebrew (Beginner)
Language Log: https://forum.language-learners.org/vie ... =15&t=9548
x 1294

Re: Here's how many words and how much time you need (probably)

Postby Beli Tsar » Tue Jul 27, 2021 4:08 pm

Le Baron wrote:^ I agree wholeheartedly with the above. Especially about the constant failure to recognise the 'human factor' when plotting graphs and running algorithms, which reeks of the delusional idea that solving technical problems means it's plain sailing for human practise. The hard fact is that you can plan a day out with what you think is military precision and it might still rain, plus you have no control over things you don't yet know and thus can't asses. Causes unnecessary frustration.

Which means a long-term goal is probably best left in outline (and essentially put to the back of your mind and humoured) whilst you get on with the actual steps of that journey.

Despite modelling things above, I too wholeheartedly agree that the human factor is more important, and more unpredictable, and that as a result solving technical problems is only a small part of the real issue. No plan survives contact with the enemy, after all. A look back at my own log forces me to realise that in a fairly humiliating way, even if I thought I knew that beforehand!

I think there is value in the modelling, even if what I did above is appallingly crude and needs to be taken with a wheelbarrow full of salt. Partly it actually demonstrates the enormity of the task: sure, the daily time commitment to 16,000 words looks more manageable than you might think here, but 1000 days of commitment is not small. And it's also a reminder that to get to this sort of level that you would need this time commitment on vocabulary alone, and indeed that even that would only be possible with a rigorous reading plan as well. Certainly anyone who thinks that they can get to a rounded C2 in 1000 days in their spare time better have a lot more time and self discipline than I do.

But when it comes to it, even if the 'human factor' is key in all these things, but that it might be easier to evaluate that if you also have a vague conception of what's necessary - and perhaps possible - through this kind of modelling. For every person who looks at the FSI numbers and makes an unrealistic study plan that they can't stick to, there must be at least one other who realises that their previous targets were unrealistic. The people asking on Reddit or whatever 'can I get fluent in three months using Duolingo before my holiday' certainly might benefit! Having a map of the terrain ahead is handy, even if I can't predict the weather - or the broken ankle - on the way.

For myself, certainly, this kind of modelling makes me realise that a slow, plodding, but determined approach is likely to get me further in the long run than an over-enthusiastic mad dash to reach a particular level.
4 x
: 0 / 50 1/2 Super Challenge - Latin Reading
: 0 / 50 1/2 Super Challenge - Latin 'Films'

User avatar
jeff_lindqvist
Black Belt - 3rd Dan
Posts: 3135
Joined: Sun Aug 16, 2015 9:52 pm
Languages: sv, en
de, es
ga, eo
---
fi, yue, ro, tp, cy, kw, pt, sk
Language Log: viewtopic.php?f=15&t=2773
x 10462

Re: Here's how many words and how much time you need (probably)

Postby jeff_lindqvist » Tue Jul 27, 2021 4:23 pm

Beli Tsar wrote:Realistically, even with one direction, your retention rate is going to be very different at different stages - low for the first thousand, then increasing rapidly as you get used to the language, and then at some point hitting much rarer words and so dropping again.


I thought it would be the opposite: better retention for the first thousand words, clearly the core vocabulary which you'd encounter all the time in any study material.
6 x
Leabhair/Greannáin léite as Gaeilge: 9 / 18
Ar an seastán oíche: Oileán an Órchiste
Duolingo - finished trees: sp/ga/de/fr/pt/it
Finnish with extra pain : 100 / 100

Llorg Blog - Wiki - Discord

User avatar
smallwhite
Black Belt - 2nd Dan
Posts: 2386
Joined: Mon Jul 06, 2015 6:55 am
Location: Hong Kong
Languages: Native: Cantonese;
Good: English, French, Spanish, Italian;
Mediocre: Mandarin, German, Swedish, Dutch.
.
x 4876

Re: Here's how many words and how much time you need (probably)

Postby smallwhite » Tue Jul 27, 2021 5:15 pm

GandwolfTheGrey wrote:Words known per level:

A1 - 500
A2 - 1000
B1 - 2000
B2 - 4000
C1 - 8000
C2 - 16000

Time needed per level:

A1 - 1 month
A2 - 2 months
B1 - 4 months
B2 - 8 months
C1 - 16 months
C2 - 32 months

As you can see, each level for both categories doubles the previous total.

Does the A2er know 1000 or 1500 words? Has s/he been learning for 2 or 3 months?
4 x
Dialang or it didn't happen.

Beli Tsar
Green Belt
Posts: 384
Joined: Mon Oct 22, 2018 3:59 pm
Languages: English (N), Ancient Greek (intermediate reading), Latin (Beginner) Farsi (Beginner), Biblical Hebrew (Beginner)
Language Log: https://forum.language-learners.org/vie ... =15&t=9548
x 1294

Re: Here's how many words and how much time you need (probably)

Postby Beli Tsar » Tue Jul 27, 2021 5:34 pm

jeff_lindqvist wrote:
Beli Tsar wrote:Realistically, even with one direction, your retention rate is going to be very different at different stages - low for the first thousand, then increasing rapidly as you get used to the language, and then at some point hitting much rarer words and so dropping again.


I thought it would be the opposite: better retention for the first thousand words, clearly the core vocabulary which you'd encounter all the time in any study material.

Interesting. Maybe I'm a bit odd here! Certainly words in met in textbooks are pretty easy to absorb. Yet I've always found things get quite a lot easier once I'm 1000 words or so in. Presumably it depends on how much of that study material you have absorbed before hitting Anki - if you are Anki-ing a frequency list, it will inevitably include a lot of words not actually in beginner textbooks.
The bigger problem I have is all the little functional words - prepositions, conjunctions, etc. etc. - that are high frequency, but hard to absorb. Beginner textbooks seem to cover fewer of these than they should relative to frequency, presumably precisely because they would overwhelm you. But when you start reading they are everywhere! They are so easy to muddle with each other. And they are often slippery things, without a direct gloss or equivalent in your own language, which is - for me - the one thing that makes a word really hard to absorb. They are the words that often get the longest discussions in lexica, after all.... I was weighing whether to buy a monograph on Ancient Greek prepositions today for this reason. You need to meet them in a lot of different sentences before you really understand the range of their use, whereas most verbs and nouns correspond to a simple idea.

Get a bit further on through a frequency list, and you're dealing with a lot more nice plain nouns, which are comparatively easy. And of course you once you hit a thousand words or so, and have been working with the language for a while, words start to seem more natural. You spot compounds, and you just have a feel for them. Thinking back to learning Farsi, I saw a substantial increase in retention and reduction in time to learn cards after this beginner stage, and didn't need to use tricks like mems/audio/pictures anywhere near as much.

But, of course, it gets hard again once you start learning words that are genuinely low-frequency.
1 x
: 0 / 50 1/2 Super Challenge - Latin Reading
: 0 / 50 1/2 Super Challenge - Latin 'Films'

User avatar
luke
Brown Belt
Posts: 1243
Joined: Fri Aug 07, 2015 9:09 pm
Languages: English (N). Spanish (intermediate), Esperanto (B1), French (intermediate but rusting)
Language Log: https://forum.language-learners.org/vie ... 15&t=16948
x 3631

Re: Here's how many words and how much time you need (probably)

Postby luke » Tue Jul 27, 2021 6:17 pm

smallwhite wrote:
GandwolfTheGrey wrote:Time needed per level:

A1 - 1 month
A2 - 2 months
B1 - 4 months
B2 - 8 months
C1 - 16 months
C2 - 32 months

Has s/he been learning for 2 or 3 months?

That's an interesting take on making the time projection algorithm more realistic. I.E., it would be more accurate to say A2 = 1 month + 2 months. But the algorithm still seems to suffer from the fact that it was created by a human (human factor) :lol:.
0 x


Return to “General Language Discussion”

Who is online

Users browsing this forum: tastyonions and 2 guests