Slow-cooked Korean

Continue or start your personal language log here, including logs for challenge participants
qeadz
Green Belt
Posts: 298
Joined: Thu Jul 21, 2016 11:37 pm
Languages: English (N), Korean (~A2)
x 400

Re: Slow-cooked Korean

Postby qeadz » Fri May 19, 2017 7:42 pm

I don't recall what I have posted to this log - I have way more ideas for things I want to say which fall by the wayside than actual posts I bother to write up. I'm also too lazy to look back in previous posts :)

So this post may or may not rehash something I mentioned before.

Recently I've been thinking a lot about vocabulary. Despite some corners of the language learning web trying to build cases for how little vocabulary one can get away with, I believe one needs to know words by the thousands to even hope of getting anywhere. I am well aware of frequency lists and how the top x number of words account for like 95% of language and such statistics...

I spent some time (a few posts back) going over the top 1000 words in a Korean frequency list and discovering that I knew 75% by their dictionary entries. I've estimated that I must know about 2000 words well enough to identify them by their dictionary entries.

Recently on Sol's log he linked an online Korean vocab test and upon doing that a couple of times, it also pinned me in the 1900 - 2200 range. So I realize they're probably making a tonne of assumptions and its highly inaccurate, never-the-less it is an external data point which adds some weight to my existing belief.

Were I to go back to when I first believed I knew around 2000 words and layer on top of that a very rough estimate of how many hours of study I have put into Korean (reading, listening, reading about grammar, writing, all included) I can then work out that I have accumulated known words at a rate of about 4 - 5 per hour.

So this brings me to two key questions:

1) If I had invested some time reviewing vocabulary (flashcards and/or wordlists) instead of just reading and listening, could I have made a significant impact on that number? I mean if I had been able to double it then I'd know 4000 words by now which would be AWESOME!

2) How is this rate of picking up known words likely to trend if I keep the status quo and simply continue to read and listen?

Question 1, I feel, probably has an honest and correct answer of: Yes, if I had to specifically do some SRS based vocabulary review then I could reasonably have achieved 6 - 9 known words per hour with even a very modest effort.

Question 2 is more interesting. I think there are multiple ways one could argue the case and each might *sound* reasonable but only one may match the reality of things.

On one hand one might consider that initially the content I was working through had very limited scope. It was beginner material. As such I was continually being exposed to the *same* unknown high-frequency words. Of course with such rampant exposure my initial uptake of vocabulary would have been quite high, but then declining as the common words became known and I was left with the less common words to learn.

On the other hand as one moves to more advanced content, the content draws upon an ever-increasing portion of the Korean lexicon (right word?) which means I am being exposed to MORE variety than ever before. Drawing upon existing knowledge, it may increase my capacity to retain new words. So perhaps I'll find my learning is beginning to snowball?

In a very general sense it seems language learners on the internet are split between the two. Basically if one subscribes to the second notion (as I believe Steve Kauffman would given what he has said in his numerous vlogs) then such a person would likely conclude that the most *useful* place for word lists is in the beginning stages of language learning - to kickstart the snowball effect.

Were one to subscribe to the first notion in which one becomes bogged down in the mass of vocabulary at non-Beginner levels, then such a person would seemingly believe targeted vocabulary review to be increasingly important as one becomes more advanced in a language.

I am feeling like some kind of targeted review might be a good thing for me going forward. So I guess the first notion is the one I subscribe to on some 'deep' level.

Just last week I was thinking about this as a very dumbed down view on what might be going on in my language learning:

If you gave me the same 5 words every day just to look at, I'd probably remember them fairly easily.

If you gave me 30 words a day just to look at, but 25 of them changed every day, I'm sure it'd take me much longer to remember the 5 recurring words.

Now as I have moved up in content, like an inverted pyramid, the strata I'm in necessarily draws on a significant portion of the Korean lexicon and so every day I see a whole stack of unknown words. So the ones which are reoccurring are clouded a lot by the ones which are much more infrequent.

I don't believe a snowball effect is happening with regard to my acquisition of vocabulary and so I am heavily leaning toward targeted vocabulary review. Maybe even taking my 2nd physical notebook which was reserved for grammar practice when the current one fills up, and instead starting to write out some wordlists - possibly based on content I am reading or possibly just following the published frequency lists.

*sigh*. My Korean is starting to move ever closer to actual 'study'. But you know if its a difference of like 2000 words over the next year, thats very very considerable. I mean would I like to be at 3000 words or 5000 in a years time?
3 x

User avatar
Ezy Ryder
Orange Belt
Posts: 146
Joined: Tue Aug 25, 2015 8:22 am
Languages: PL (rusting Native)
EN (Advanced)
中文 (Lower Intermediate)
日本語 (Beginner, not studying)
台語 (Dabbling)
Language Log: viewtopic.php?t=1164
x 214
Contact:

Re: Slow-cooked Korean

Postby Ezy Ryder » Sat May 20, 2017 10:41 am

Did somebody say "vocabulary"/"SRS"? ^^
qeadz wrote:1) If I had invested some time reviewing vocabulary (flashcards and/or wordlists) instead of just reading and listening, could I have made a significant impact on that number? I mean if I had been able to double it then I'd know 4000 words by now which would be AWESOME!

I'm absolutely positive SRS is capable of making a significant impact on that number. In fact, I'd say, with SRS we're speaking of dozens of words per hour.
qeadz wrote:2) How is this rate of picking up known words likely to trend if I keep the status quo and simply continue to read and listen?

I haven't studied English vocab much the last few years, yet I can see improvements both on my own, as well as in the estimates of some online tests. I think it was like 2-3k words a year (just an estimate, might be very inaccurate), and I essentially use English more than any other language. So, you'll definitely keep improving, so long as you expose yourself to adequate amounts of comprehensible input (although I don't know how comprehensible what you've been reading/listening to so far was).
qeadz wrote:Question 1, I feel, probably has an honest and correct answer of: Yes, if I had to specifically do some SRS based vocabulary review then I could reasonably have achieved 6 - 9 known words per hour with even a very modest effort.

It's a gross underestimation.


qeadz wrote:*sigh*. My Korean is starting to move ever closer to actual 'study'. But you know if its a difference of like 2000 words over the next year, thats very very considerable. I mean would I like to be at 3000 words or 5000 in a years time?

That's an important question you have to ask yourself. Is what you've been doing so far fun/enjoyable enough to make up for how much longer it's gonna take to get to a level you find satisfactory? Or is the fun you could have at a higher level worth the less enjoyable effort of "studying" to get there sooner?

Here's one way you can try to estimate how much time a day studying x words a day would (eventually) take: In Anki (with the default settings), you'll be seeing each new word 2-3 times on the first day. The average number of reviews (provided you finish all your reviews daily, and learn the same number of new words daily) eventually gets to 11-15*x (your mileage might vary. The degree of "unfamiliarity" withe the material, as well as the leech threshold can affect this). Assuming the typical 1 in ~5 failure rate, you'll be seeing ~80% of the words you review once, and (with the default settings) the remaining ~20% about twice.
All in all, that estimate would come down to about 16.2*x*t to 21*x*t (where t=average time per review). Some example values @ 6s per review:
1,826.25 words a year (5 words a day) = 8m06s to 10m30s a day
3,652.5 words a year (10 words a day) = 16m12s to 21m a day
10,957.5 words a year (30 words a day) = 48m36s to 1h03m a day

PS SRS'ing doesn't mean you can't do more enjoyable/other things on top of it. In fact, many would argue it's best if you SRS on top of doing more enjoyable/other stuff.
1 x
阿波
: 1250 / 1000010k SRS Challenge :
: 3750 / 48084,808 漢字 (handwriting) :

AndyMeg
Blue Belt
Posts: 633
Joined: Sat Jan 28, 2017 2:44 pm
Languages: Spanish (N), English (B2-C1), Japanese (A2-B1), Korean (Lower Intermediate?)
Language Log: https://forum.language-learners.org/vie ... 02#p201902
x 1302

Re: Slow-cooked Korean

Postby AndyMeg » Sat May 20, 2017 1:22 pm

qeadz wrote:I don't believe a snowball effect is happening with regard to my acquisition of vocabulary and so I am heavily leaning toward targeted vocabulary review. Maybe even taking my 2nd physical notebook which was reserved for grammar practice when the current one fills up, and instead starting to write out some wordlists - possibly based on content I am reading or possibly just following the published frequency lists.

*sigh*. My Korean is starting to move ever closer to actual 'study'. But you know if its a difference of like 2000 words over the next year, thats very very considerable. I mean would I like to be at 3000 words or 5000 in a years time?


If you plan to make your own wordlists, I would suggest that you try to choose the vocabulary wisely. Some people just put in there each and every new word they find, but imho I think time is best spent if from those new words you first focus only on the most frequent ones. If you want to know which are the most frequent ones, the sense of familiarity may help: in my case, for example, if a word is high-frequency, then I usually have a feeling that I've seen/hear it many times before.

I don't know if it'll be the same for you, but I've noticed that I remember words better if I study them grouped by categories (usually "part of speech": nouns, verbs, adjectives, conjunctions, adverbs, pronouns, endings, etc. But also some other kind of categories: like numbers, parts of the body, etc.)

I think a good balance between hard work and enjoyment is vital to keep motivation high. So I suggest that you try to find ways in which your study time feels as enjoyable as possible.
1 x
Beyond The Story 10 Year Record of BTS Korean version: 36 / 522

qeadz
Green Belt
Posts: 298
Joined: Thu Jul 21, 2016 11:37 pm
Languages: English (N), Korean (~A2)
x 400

Re: Slow-cooked Korean

Postby qeadz » Tue May 23, 2017 5:06 pm

Ezy Ryder wrote: ... SRS related championing ...


Were I to take up doing SRS revision, word lists or some kind of vocabulary review, it would only be for closer to 5 - 10 minutes a day. Basically my plan would be to retain my current 1 hr of Korean as-is but add on some small bit of vocabulary review.

So looking at your napkin math it seems I'd be correct in my original estimate that with this tweak I could add an additional 2000 words to my vocabulary over the course of the next year (I'll assume 365 hours).

In fact the prospects of this kind of increase in vocab is so alluring that I daresay I have already convinced myself of this decision even though I pretend I have not made it yet!

So then the question is: If I have already made up my mind, what am I waiting for?
Last edited by qeadz on Tue May 23, 2017 6:15 pm, edited 1 time in total.
0 x

qeadz
Green Belt
Posts: 298
Joined: Thu Jul 21, 2016 11:37 pm
Languages: English (N), Korean (~A2)
x 400

Re: Slow-cooked Korean

Postby qeadz » Tue May 23, 2017 6:13 pm

AndyMeg wrote:If you plan to make your own wordlists, I would suggest that you try to choose the vocabulary wisely.


Oh, for sure! That is good advice. I think word frequency lists will probably play into whatever I do going forward.

I mean eventually I want to know ALL the words but since I can only learn so many at a time, I should start with the most useful.
1 x

AndyMeg
Blue Belt
Posts: 633
Joined: Sat Jan 28, 2017 2:44 pm
Languages: Spanish (N), English (B2-C1), Japanese (A2-B1), Korean (Lower Intermediate?)
Language Log: https://forum.language-learners.org/vie ... 02#p201902
x 1302

Re: Slow-cooked Korean

Postby AndyMeg » Tue May 23, 2017 8:41 pm

qeadz wrote:So then the question is: If I have already made up my mind, what am I waiting for?

Maybe you are waiting to find the right SRS app for you?

Personally, I don't like using Anki, but lots of people love it. I usually prefer Memrise (it is more engaging to me). One course I'm actually enjoying a lot is this one: Every TTMIK Lesson (Levels 1-10)

qeadz wrote:I mean eventually I want to know ALL the words but since I can only learn so many at a time, I should start with the most useful.

I think the general lists of more frequent vocabulary are best until the 4.000/5.000 or even the 10.000 words range. After that I think that some area/topic-especific frequency lists are better, or maybe just being in contact with lots a lots of native material with some time dedicated to things like "intensive" reading or some other kind of "intensive" practice. At that stage you could target different areas in sequential order. For example, if you like suspense novels, you could try to read as many suspense novels as you can (combining extensive and intensive reading), and then target another particular area of interest.

For frequency lists it is also important to know the corpus (the data from which the frequency was analized). If you use a frequency list with a corpus of only newspapers, then you'll get a jumpstart in reading newspapers, but maybe find some difficulties with the most common vocabulary in other areas like TV, dialogues or podcasts.
1 x
Beyond The Story 10 Year Record of BTS Korean version: 36 / 522

Sayonaroo
Green Belt
Posts: 256
Joined: Wed Jan 11, 2017 12:13 am
Languages: English(N), Japanese -fluent?, Korean - advanced?, Spanish (b1?)
Language Log: http://choronghi.wordpress.com
x 319
Contact:

Re: Slow-cooked Korean

Postby Sayonaroo » Wed May 24, 2017 12:23 am

qeadz wrote:
Ezy Ryder wrote: ... SRS related championing ...


Were I to take up doing SRS revision, word lists or some kind of vocabulary review, it would only be for closer to 5 - 10 minutes a day. Basically my plan would be to retain my current 1 hr of Korean as-is but add on some small bit of vocabulary review.

So looking at your napkin math it seems I'd be correct in my original estimate that with this tweak I could add an additional 2000 words to my vocabulary over the course of the next year (I'll assume 365 hours).

In fact the prospects of this kind of increase in vocab is so alluring that I daresay I have already convinced myself of this decision even though I pretend I have not made it yet!

So then the question is: If I have already made up my mind, what am I waiting for?


well perhaps if you SRS correctly. of course everyone has their own preferences but if you're thinking of doing a pre-made deck of some sort it may not work out. If you fail/forget cards the reviews will pile up (it's more likely with a deck you did not make especially if it's made from a textbook etc etc with bland/sterile sentences) and also it depends on what format you set up your anki deck. I recommend making your deck. ESPECIALLY for Korean all the "good" "awesome" words are not on the frequency lists... you will see what I mean if you start mining words/sentences from native media

you wrote that you're contemplating whether you'd have gotten further if you ankied sooner but I wouldn't worry about it because at least you didn't waste time anking stuff you can't remember (having a deck hundreds of reviews piled up or a lot of failed cards). Currently, I'm learning spanish leisurely and I do not use anki very much (I have made some sub2srs cards) because I know it's pointless since I can't remember Spanish words as easily as korean or Japanese words. In my opinion anki is not useful for beginner/intermediate stages because you're better doing other stuff with your time.
1 x

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: Slow-cooked Korean

Postby smallwhite » Wed May 24, 2017 4:35 am

In contrast to Anki users who use it but dread it, Quizlet users use it and love it. See my thread here:
Do You Like Quizlet?
1 x
Dialang or it didn't happen.

qeadz
Green Belt
Posts: 298
Joined: Thu Jul 21, 2016 11:37 pm
Languages: English (N), Korean (~A2)
x 400

Re: Slow-cooked Korean

Postby qeadz » Wed May 24, 2017 6:09 pm

Actually I'm not a stranger to flashcards - years ago I used Memrise for some time. I didn't have a bad experience with it, but on the whole it wasn't terribly effective I don't think. Certainly last year's Korean study which was primarily reading and listening far surpassed what I achieved with my time doing Memrise.

When I did Memrise, it was with a pre-made frequency list.

Now basically theres nothing magical about flashcards or SRS really. We know that 'calendar time' plays a role in building up those mental connections in addition to 'study time' (total amount of time spent on a word or phrase). SRS is simply paring down the 'study time' to the minimum by making best use of 'calendar time'.

The flashcard itself is just the paring down of information. In a regular environment a word or phrase occurs in an article. So why read the whole article again? Just pick out the specific word/phrase and do targeted review on it.

I'm under no illusion that an alternate solution to my problem (being: I want to retain more vocab) would be to select articles with a certain portion of unknown words and reread them in an SRS fashion. Right now I am probably not rereading articles enough and certainly not optimizing 'calendar time' when I do reread them.

So there are a lot of questions actually. Some have very simple answers (IMHO) and some are debatable.

Why return to SRS?
Because with limited time and broad exposure I *feel* I'm losing effectiveness in being able to commit words to memory.

Why return to isolated review? (flashcards or wordlists)
Because a lot of time is spent reading or listening to an entire article, and while there is good benefit to re-reading, I also crave novelty. So 'isolated' review will let me cover more words which need reviewing in a shorter time, leaving more time left over for reading new material.

What went wrong last time I used Memrise?
Well the words I found I remembered best at that time were ones for which I had some kind of mnemonic. However it was not possible to come up with non-overlapping mnemonics for everything. So the majority of words were simply (to my ear) a collection of random syllables whose meaning I had no immediate use for.

Pre-made or Custom 'deck'?
Recently I discovered that there were a not-insignificant portion of the 'top 1000' words in Korean which I have absolutely no recollection of. Even seeing the English translation didn't help - I'm pretty sure I have not encountered the word anywhere. So @AndyMeg is correct that the sources for the frequency lists makes a substantial difference.
I don't want to be learning words I don't expect to encounter in the material I read. Eventually I want to know those words too, but I'm a big believer in feedback mechanisms. So I want to read it, hear it, write about it, review it - get that word showing up in a number of places.
To this end I think making my own custom list is the way to go - adding words I come across in the material I read.

Maybe I'll write a little program which, if given a text in Korean, finds the dictionary form of all the words (if possible) then filters them through a 6000 word frequency list and adds them to a deck for review...

Anki, Memrise, Quizlet or Word lists?
I've only used Memrise. I have not used the others so I shall take a look at what they will involve.
3 x

AndyMeg
Blue Belt
Posts: 633
Joined: Sat Jan 28, 2017 2:44 pm
Languages: Spanish (N), English (B2-C1), Japanese (A2-B1), Korean (Lower Intermediate?)
Language Log: https://forum.language-learners.org/vie ... 02#p201902
x 1302

Re: Slow-cooked Korean

Postby AndyMeg » Thu May 25, 2017 2:29 am

I'm currently using Memrise's android app. For most of the words (or sentences, in the case of the TTMIK course) I don't use the mnemonics. What I like about the app is the variety of test modes they use: sometimes I have to choose the right korean syllables in the right order to form the word/phrase/sentence they are asking for, sometimes I have a multiple choice test, sometimes it is a speed game-test for which I have three lives, and there are other test modes too. For now I'm on the free version, but from time to time it lets me "try" the paid modes (like difficult words, only audio, etc.).
0 x
Beyond The Story 10 Year Record of BTS Korean version: 36 / 522


Return to “Language logs”

Who is online

Users browsing this forum: Henkkles, Link and 2 guests