Tired of Anki? Try Goldlist

When I get done with my hybrid experiment I'll have some numbers. There is the 2nd day Iversen list review and the subsequent day first real Anki review, and I have been logging my miss rates for same. So yes I could pass a vocab test, but only getting somewhere between 70-85%. I suck at mnemonic associations when there is not an obvious one suggested by the morphology.

Hi, another HTLAL-er here. I recognise loads of names from there of course, and maybe some of you recognise me. Been meaning to join for a while but never got a round tuit. But I wanted to join in a current conversation on Goldlists, since I'm revisiting it myself.

Bob and others have explained the system very well, so not much for me to add. I was originally a little unclear on whether you were actually supposed to go beyond the "silver book" level (which would take you to the 4th, 5th, 6th, and 7th distillations), or not. I'd sort of got the impression that the "gold list" was more of a concept than a real object. But now having re-watched all his relevant videos, as someone here mentioned, he does actually use a "gold list" book himself, although shared between more than one language.

I think the 2-month upper-limit suggestion is just that, and is only supposed to help you keep up your momentum. He says that he himself has gone back to lists that were 6 years old, although in that case he would remember much less than 30%. (Although that seems to contradict his point that about 30% should have gone into long-term memory doesn't it?).

I must say that 10 headlists a day sounds far too intensive (for me), and at the moment, I'm only doing one (or maybe two at weekends), That sounds very modest, but I wanted to keep it very simple, and even that that modest rate, you are seeing over a thousand items in 6 weeks, and 9125 in a full year, which is not so dusty. (You'd need at least another 12 weeks to finish silver-listing the last ones of that full year, so it would take about 1 year and 3 month to complete 9125 items, although of course you could be starting new headlists in new "bronze books" in that 12 weeks, if you wanted).

But using Prof. Arguelles criteria, you'd easily then have enough words to be extensively reading at the 98% comprehension level (I think he says that's about 8-9000 words for English anyway).

Speaking of the Professor, someone posted a link somewhere in this forum a while back to his 2 videos on selecting material for extensive reading, which are fascinating.

Elsewhere, the professor has said that in the early stages of a language, he never uses dictionaries, but uses parallel translations or translations to quickly identify unknown words, in a similar way to the way Davey describes using translations for the same purpose, in order to get words to put in your Goldlist headlists.

Now, somewhere, the prof. says that while for conversation, you can be at 98-100% comprehension level with "only" 6-7000 words, to get to 98% comprehension level for reading you need 8-9000 words (at least in English). He further makes the point that since unusual words crop up only rarely in speech, the best place to learn them is by reading. Obvious when you think, but I'd never though about in quite that way.

Although I have Goldlisted before in German, in a way, it wasn't a good test, because I'd been doing German a long time before, and tended to have seen a lot of the words many times.

However, I've now started using Goldlist with my Welsh, having completed the beginner and intermediate SSiW courses (and some of the advanced). Those courses are great for getting you speaking, but do not claim to give you a large vocabulary. This is actually pretty much in keeping with Uncle Davey's advice anyway: he suggests using an audio-only course in the beginning (like Michel Thomas or Pimsleur) to learn pronunciation, etc, before starting Goldlisting.

I'm not sure if it's been mentioned explicitly in this thread, but I think the key idea behind his view of the way the GL is supposed to work is that you put no effort whatsoever in trying to learn the words. You just basically "show them" to the subconscious and then let it do the work.

This I would think is the main difference between GL and Iversen's word-lists, since with the latter you are encouraged to "memorise" the words, although the specifics of how you do that are left to you. Iversen says (more or less, apols if I'm wrong in details) that the item can be verified to have got from short-term memory into long-term memory in 24-48 hours, while Davey says you have to wait 14 days or more to be sure. I think Iversen is probably closer to the currently-known science of these things, but nevertheless, I'm intrigued to see if Davey has hit on something not quite known to science yet.

I don't take too seriously his claim that his work is based on Ebbinghaus, to be honest, and in a way, I don't care. If it works, that's good enough for now, and maybe, exactly why it works will one day be known (and maybe it can thereby be enhanced.

Some further thoughts:

In the comments to Viktor's blog posting about GL, he quotes Davey as saying it takes on average 1 hour to learn overall an initial headlist of 25 words (this learning being done over time in several iterations of course). Taking that as a guide, you could therefore learn 10,000 words (to take a nice round number) in 400 hours on average, which you could do in a year at the rate of about 1 hour 10 minutes work per day. That sounds quite impressive if it works, and if one could be so consistent. You'd easily have enough comprehension for conversation, and also enough for extensive reading, except perhaps for the most difficult books (such as Moby Dick in English, to use the Prof's example).

I'm not quite sure if I understand what people mean in earlier posts by "easy" words, but I suppose in Welsh it would be words that are obviously borrowed or at least heavily cognate with English. For example, this morning I came across "ffeil", meaning "file" (the sort you put papers in, I think). It looked unfamiliar, until I realised what it was, and it's pronounced basically the same as English, but with a Welsh accent. There are a lot of words like that, and I would not bother to GL them.

Where I'm getting most of my GL words for the moment is a simple book in the "Novelau Nawr" series, which only has about 80 pages, and includes vocabulary footnotes and end-notes. It's quite an enjoyable read and is in modern Welsh, close to the spoken language. I think I will continue with this sort of book for quite some time, as they are fun, but also useful and cheap on Am. marketplace. I also get odd words from internet radio and TV and the web, and also my learner's dictionary (but only words I think I will come across). The great advantage of getting words from books is that you see them used in context and see how they are used, e.g. how they are affected by mutation, which is quite important in Welsh.

At my current (modest) rate, it's going to take me from 2 to 3 months to fill up my first "bronze" book, by which time I will have some idea of whether it's working for me. Festina lente.

EDIT: to add some thoughts triggered by re-reading the whole thread:

Davey's article about scheduling: I lost the will to live quite soon after starting to read that... No, actually it was quite interesting, but there is no way I'd approach it in that complicated way. You can see how it would appeal to an accountant though.

You are supposed not to work for more than 20 minutes, and have at least a 10 minute break between stints of work. I break this rule because it takes me more than 20 minutes to complete a headlist, and I like to finish it in one go. But I do it at a leisurely pace. Since I'm mostly only doing 1 headlist per day, breaks aren't much of an issue yet, but will be when I get to the distillation stage, which won't be for a week or so. I may have to revise my thoughts once I get going with that.

For me, since I'm going through wordlists and dictionaries, it's usually closer to 15 mintues for me. I feel there's also a sort of distillation even when just headlisting, because I'll throw out easy words, or words that I feel I could figure out to get a "good" 25. I've also noticed that when I'm working with a closely related language, I'm usually able to take out allot more than 1/3 by 4th distillation.

It does not seem unreasonable to bump the time up to 30 minutes, based on other marerial that I've read on how to study. You also could stop in the middle of a list, and come back to it after a break. It doesn't really make a difference. I think the important thing is to keep yourself from feeling overwhelmed or that you are "forcing" yourself to keep going.

I've taken up listing more hebrew at a lesuirely pace. It felt like more of a chore the first time around, compared to other languages. I probably broke the 20 minute rule often. But now after reading through Jonah it seems easier now. Maybe i've got a good "core" to work from after reading that thing so many times.

Diolch (thank you) Bob. I think I went for only 1 headlist a day, because I was afraid of getting overwhelmed later on at the distillation stage. I've shied away from Anki for fear of becoming a slave to it. With GL and wordlists, you can work at your own pace essentially, but with GL, knowing that there is a distillation process coming up later, I could easily feel overwhelmed with multiple headlists per day, but perhaps I'm being overcautious. But it would be silly to choose a low-stress method and then let it become stressful.

On vocabulary tests, I feel that Uncle Davey would say these are contrary to the spirit of GL, although on the other hand, I think it's only human nature to want to test yourself from time to time.

I think my criteria for including words in headlists are roughly:

- Do I know it already? - reject
- Do I half know it? - reject
- Is it an obvious loan-word or very similar to English? - reject
- Is it something I really think I am not likely to need to know? - reject
- Is it something I might need in conversations? - accept
- Is it something I won't need in conversation, but would want to knowing reading or listening? - accept
- if in doubt - accept.

Davey talks about "project size" in his article about scheduling. Well, I don't know in advance how many words I'll want to know - it's a bit open ended, but my gut feeling is I want to get up into the Prof.'s "extended reading" - 98% level so I think I'm ultimately aiming for at least the 9-10,000 mark, which actually to me sounds staggeringly ambitious, but who knows. Theoretically it's doable in not much more than a year. We'll see.

I think when I run out of those learner's novels I mentioned, I still won't be anywhere near that number and it may be below 5,000 or much less, so I 'll have to move on to more advanced novels, which is actually a problem, since there aren't that many translations available so far as I can tell, and even fewer parallel ones in Welsh. I don't want to be referring to dictionaries all the time, even e-dictionaries. I suppose e-books and pop-up dictionaries will have to be the answer, although I'm slightly loath to be dependent on electronic tools.

Diolch yn fawr iawn! (Thank you very much!).
And thanks for setting up this nice forum.

This is a response I got a long time ago from Victor Berrjod over at HTLAL about how he does Chinese characters using Goldlist. I thought it would be useful to add this to the thread.

Since I've had some success with Goldlisting Japanese (based on the last post) I thought I would add what I did here. After doing some, I recognized some Kanji on Naruto (of the different factions) and what they meant, and then since they didn't translate the titles of the leaders, the roots of the word were right there in the subtitles and I soon heard them saying these words. Now that's cool.

I started off Goldlisting the Hirigana, and when that was done, I did a two line thing for vocab.

I didn't list Kanji, I listed words. If you don't know, in Japanese, sometimes a Kanji is a whole word, but often times not, and it needs some trailing Hiragana, So I listed the whole word. Once with the pronunciation in "romanji", another time with a translation. If it's just Hiragana, I list both things one the same line, as it's still shaky and I haven't really used it yet. Katakana I treat like the Hiragana and it seems to be working fine.. I'm starting to pick up on that alphabet as well.

When I take out "half" a kanji based word, I rewrite all the information on one line, as it's conceivable that you might not automatically recognize the half you checked off the first time. More specifically, It seems wrong somehow to have a session where you try to remember how to pronounce a word, but not what it means.

I'm routinely taking out 13 lines the first time around instead of the usual 8. The percent stayed high on second distillations as well. I guess it's to be expected because I'm doing allot of double listing initially.

In retrospect I'm not sure the Hiragana listing was needed, as I'd probably pick it up in the pronunciation lines anyways. I'm already picking up on the diacritics that I ignored

Someone else's experience... http://www.isimplylovelanguages.com/review-goldlisting/