How to (not) learn a language in 12 hours


My friend D asked me a good question today. Why would someone want to spend 12 hours straight studying a new language?

That’s what two other language enthusiasts and I did on Saturday, studying Bahasa Indonesia from 9am until 7pm, and then speaking it with each other for two more hours. The precondition for the “12-hour challenge” is that none of us could know or learn a single word of the language ahead of time; another was that we choose a language we were all actually interested in learning. Bahasa Indonesia was one of two languages that fit both criteria.

The challenge was inspired by a video showing four polyglots picking up Romanian in just one hour. The results are pretty amazing. After just an hour with a Romanian tutor, they are able to hold a somewhat fluent conversation. My language friends and I weren’t confident we would fare as well in only an hour, but we thought we could get close if we were given ten times that allotment.

Sportingly, I didn’t think much ahead of time about how I was going to spend my 12 hours studying, let alone learn anything about Bahasa Indonesia beforehand or look up any resources. Unsurprisingly, I spent the first few hours stuck somewhere between confusion and panic.

How would you start learning a new language, if your goal was to maximize your conversational ability in 12 hours? Here’s what I did, with hindsight enabling me to give a concise description that belies the chaos I was surrounded by.

Method #1: the scattershot

I started with a scattershot approach, watching some of the IndonesianPod101 3-minute lessons, trying and quickly dropping a Memrise deck, downloading some pre-made Anki decks (there is actually a pre-sorted sentence deck with audio for Bahasa Indonesia in the same vein as the ones I used for a while to learn French and German), and browsing the audio lesson library at IndonesianPod101. I watched videos and shadowed podcast lesson dialogues for a couple hours before realizing this was a hopeless strategy. There simply wasn’t enough time to learn the kind of volume of material I would need to develop a natural, acquisition-like sense of the language.

I grudgingly realized this kind of language learning necessarily resembled cramming for a test much more than it resembled acquiring a language. Clearly I needed to memorize some more useful phrases, the kind of thing I would be using that very night at our Indonesian table.

Method #2: the dialogue

I figured a dialogue would be the easiest package to consume these phrases in, since stories are more easily memorized than individual sentences. I wrote a dialogue of 13 lines in English, and then used Google Translate to turn it into Indonesian, cross-checking the Indonesian-to-English translation to make sure I wasn’t way off. I knew this would probably result in some unnatural phrases, but I decided to lower my standards for the sake of expediency.

How to memorize the dialogue? I turned it into an MCD-style Anki deck, but then found I had another problem: I didn’t know how to pronounce the words I was trying to learn. Bahasa Indonesia is written with the Latin alphabet, but of course the phonetics are completely different from English. I had thought I would get around this by picking up the sounds from my shadowing exercises, but since I wasn’t shadowing anymore I needed to find another workaround.

I used Google Translate’s text-to-speech feature to get an idea of the pronunciation, and then recorded myself doing an imitation of the Google Translate voice, and put the audio from my own imitation on the Anki cards, to accompany the text.

It probably should have been obvious before, but it was only around this point that it started to sink in that this method wasn’t going to be a winner either. Not least because, even if at great effort I did memorize this dialogue perfectly, I wasn’t going to get any closer to fluency in Indonesian than I got in German after memorizing my very first Slow German dialogue. Which is to say still far, far away.

Method #3: the combinatoric

In a last-ditch effort, I decided to try and enumerate all the words I would need to hold a basic conversation, covering all the parts of speech, and just memorize as many as I could. This was an even further departure from what I’m used to with language acquisition.

I made a list of personal pronouns and a list of verbs, and used a few online dictionaries to get the Indonesian words. The verbs included: to eat, to drink, to go, to want, to like, to live, to speak, to think, to know, to have, to buy, to be able, to give, to make, to tell, and to be. I tried making a list of nouns but couldn’t get inspired. I forgot all about adjectives.

Then I got cold feet and instead of persevering with individual vocabulary words I made up sentences using each of the above verbs, thinking this would be 1) easier to memorize than individual words, and 2) would help me incorporate some other parts of speech without having to come up with any more long lists. Did I think to look up frequency lists for individual parts of speech? No, for some reason I did not.

I also forgot until much later useful things like prepositions, impersonal pronouns (this, that, who, what, enough), common adverbs (how, when, where, why, also, too), super common time-related adjectives and nouns (next, last, yesterday, today, tomorrow, day, month, week). There were plenty that it never occurred to me to look up at all, and it was actually kind of fun figuring these out from context later on, during our Indonesian conversation.

Incomplete list of words and phrases in hand, I set about making Anki cards for them all. Unfortunately, at that point I didn’t have enough time to make any meaningful headway memorizing the cards (digression, but I should mention that this isn’t even how Anki is supposed to work. The whole point of an SRS is for memories to build up bit by bit, strengthening over a period of days, weeks, and months. I just don’t know a better way to memorize a lot of information quickly).

The showdown

I showed up for our conversation at the cafe feeling embarrassingly unprepared. Luckily, though I may not have memorized many words, I had all my notes, and I spent the two hours referring to them obsessively and fleshing them out when I encountered a new word from one of the other two.

This part, where we conversed using the words we had learned, was the most fun part of the day for me. We were originally supposed to have a native Indonesian speaker at the table with us, but we were informed that morning that he was out of town. As a result, there was no one at the table but the three of us stumbling through our astonishingly rudimentary conversation with abundant hand gestures and occasional pointing at our notes or showing pictures from Google search results.

It’s certain we were speaking something a bit different from what a real Bahasa Indonesia speaker would recognize as his or her own language, but I think it’s cool that we were still able to communicate. We talked about what we had done during our respective study time, we talked about our ages, what time we went to sleep, and whether we would keep studying Indonesian, and nobody spoke a word of English or any other language that we all knew.

The point of it all

What did I learn from all this? First of all, it’s pretty cool to know that I’m only one long plane ride’s length of time away from basic competency in a new language. In that sense the activity was useful as a proof of concept.

There’s one glaring omission from my day of learning Bahasa Indonesia: I never practiced with any native speakers. I could have gone on Italki or any number of other websites and found friends or tutors willing to speak to me — for money if necessary, but even then probably not a lot. Why didn’t I? At first I used the excuse that I wasn’t quite ready, that it would be more effective to learn some basics before I started speaking. And then I just got carried away with what I was doing and didn’t think about it again.

I may do this type of language marathon again someday (the other candidate we considered was Thai), and if I do, it would be interesting to try the opposite strategy: only speaking, for the whole ten hours of preparation. Now that I’ve been through a two-hour Indonesian conversation with practically no preparation (ten hours of ineffectual stressing out doesn’t actually count, does it?), the idea of jumping into the deep end again isn’t quite as scary. I would probably want to have some paid tutors lined up for the first few hours to teach me the basics, and then I could try to inflict my poor language ability on some hapless conversation partners. This approach would have an extra advantage with a language like Thai, where I wouldn’t need to worry about learning the writing system to decipher written teaching material (though taking notes might involve some ad-hoc transliteration, which could be tricky).

What else? Getting a sense of the set of words and phrases that I actually used in my first Bahasa Indonesia conversation actually gave me a better sense of what’s missing from my Korean vocabulary (and German too, for that matter). It’s inspired me to make more flashcards for Korean with things like the numbers, times, basic verbs, prepositions, etc., which might make me able to start having more fun conversations in Korean soon. This is actually what I was about to do tonight before I got sidetracked into writing this long blog post.






My name is Isaac.      이름 Isaac 입니다

What’s your name?     이름 입니다?

I’m 29 years old.      이십 살 입니다

How old are you?      살입니까?

I like to learn languages.      언어를 배우고 싶습니다.                         

I’ve learned Korean for two months.     나는 개월 한국어를 공부했다

I also like to cook, and bake bread.      요리 빵을구워 도싶습니다.      

I work in a bakery in Taiwan.     대만의 빵집 에서 일한다

What do you like to do?     취미 는 뭐니?

This is what happens when you don’t go to school

I’ve been learning Mandarin Chinese for about eight years, and I’ve been told many times that I have an almost native-sounding accent. The people who said this were doubtless exaggerating to pay me a compliment, but it’s been enough to make me complacent.

The problem is, after hearing me speak Chinese for a while, it soon becomes apparent that something is wrong. I have a good accent, and I can banter almost like a Taiwanese person (huge gaps in cultural context/awareness notwithstanding, but that’s another topic for another time), but where I fail is in differentiating levels of formality. When things get serious and I have to stand up and make a speech, or open a bank account, or negotiate with my gym’s sales representative — all situations that require a more formal tone — I still just do the banter.

I’m like a Taiwanese person who never went to school. This is not to say that I’m illiterate. But almost all of my reading and writing has been done in the context of personal communication. My reading hasn’t progressed much beyond manga. I can count the number of novels I’ve read in Chinese on one hand. And aside from a few months in Chengdu when I was just starting, I never studied Chinese in a classroom. It’s no wonder I sound uneducated.

Maybe this realization will finally get me to sign up for Chinese classes. In the meantime, I’m just going to try reading and writing more in Chinese, on the assumption that this will also improve my speaking. Specifically, I’ll start making some entries here in Chinese, and then I’ll ask a native speaker to correct my writing as if correcting a paper in school, using more formal writing standards.

Wish me luck!

January: Korean dialogues, Indonesian marathon

Pitfalls from lack of context

Last month I mentioned how I asked a Korean friend to record some simple, extremely common phrases for me, and how I was listening to them and shadowing them every day. I took it a step further and even made audio flashcards out of them, putting the Korean audio on one side and the Korean transcript and English audio on the other. Unfortunately, after a few days of reps this backfired and I completely lost motivation for a couple weeks, and then finally deleted the cards.

Why was this such a failure? First, I think putting the English audio on the card was a mistake. Putting any English on the cards isn’t ideal, but it seems like somehow having to listen to the English audio recordings when I did the reps was short-circuiting my incipient Korean brain and my English brain, and made it feel like I was getting yanked out of Korea World every time I heard them.

Second, I think it was a mistake to ask for these sentences out of context. Listening to a long list of simple phrases is kind of confusing, and also kind of boring, and I think this prevented the phrases from actually sticking even when I listened to them hundreds of times. My passive listening brain just tuned them out. Contrast this with the level one dialogue from Talk to Me in Korean, which I can still pretty faithfully repeat from memory even though I haven’t listened to it in weeks.

Again, this proves the importance of stories.

Speaking of Talk to Me in Korean, I’ve still been listening to these audio lessons sporadically over the past month. They’re entertaining enough, and what makes them really valuable to me is that I’ve been using this audio comprehension deck that takes the audio directly from the lessons. The lessons help me understand the cards (and give me stronger memories/impressions that make them stick), and the cards reinforce the lessons.

Given the high quality of the material, I really wish they would produce more dialogues. If there was a dialogue showcasing usage of the words or grammar points from each lesson, that would be so useful.

Korean dialogues: Korean Class 101

I got back to Taiwan a couple days ago and this gave me an occasion to try to get back on track with Korean. I found another promising resource: Korean Class 101. I was skeptical of it at first because of how commercial it is, but I think I dismissed it too quickly: it has tons of Korean dialogues of all levels available for download, they come with transcripts (and, if needed, lessons explaining the contents in English), and the voice actors do a good job. The key is to not get bogged down listening to the lessons themselves.

I put a bunch of the dialogue mp3s on my phone for convenient shadowing. The method I’m trying first is 1) listen to the dialogue once, 2) read the transcript or listen to the lesson as necessary to understand anything I didn’t catch the first time, 3) shadow ~5 times a day. At this level the dialogues are all less than 30 seconds long, so I can do several a day and still spend under 15 minutes. I’m started with the very lowest level, which might be too low; I’m wondering if I’ll get greater returns and have more fun if I aim a little higher.

Next language: Indonesian

Some language friends have talked me into trying a 12-hour Indonesian-learning marathon this month, on Saturday the 21st. I’ve been skeptical of these, because it seems like too short a time to actually acquire anything that will last. But these friends have more language learning experience that I do, so it seems likely I’ll learn some interesting things by going along with their plan. Maybe they’ll prove me wrong.

Doing this also gives me an excuse to keep focusing on Korean and German until the 21st, and helps cover up the fact that I neglected to start another language before and it’s now already the middle of January (I was still in the US until the 12th, give me a break!).