About half a year ago, I wrote about our experiences living in France for a year. One of the major themes was about the difficulties of learning a new language. At the time, I hypothesized that the best method of language learning I had found was one-on-one Skype based tutoring and that in three months of just using that method, I would be as good at Japanese as I was at French.
This hypothesis was incorrect.
This post will go over how well I did in my challenge, why I think it fell short of my theory, my new revised theory and what my next language challenge will be.
What I Mean By “Language Learning”
Before I get into my experiment and my opinions on language learning, I should clarify what I mean by the phrase.
First of all, I don’t know any language other than English at a conversational level (yet). I can and have had very basic conversations in Spanish, American Sign Language and French. I am probably hovering around an A2 level in each of them.
When I talk about learning a language, I mean gaining the skill to understand it spoken at native speed by native speakers and being able to respond in any way that is understood. I don’t much care about perfect grammar, I don’t much care about how well you can write it, I care about communication with humans in real life. Can you transfer and receive information with them or can you not?
Just about everyone I know has taken years of language classes in high school or via computer programs and has no confidence or ability to speak or understand it in nearly any real situation. It is not language learning to me if you can conjugate tables of verbs but can’t ask for directions. This experiment is about how to communicate and understand spoken language in real world scenarios as quickly as possible. I feel like you can improve your grammar after you understand things.
My method was to Skype with real native speakers as often as possible in the hope that in a short time I would be able to speak and understand the language in real scenarios.
How Well It Worked
I was pretty consistent and disciplined with my studies. My first class was on July 27th and my last class was on December 3rd, right before an actual trip to Japan. Nearly exactly 4 months of studying. There are roughly 90 weekdays between those dates and I completed 73 one hour Japanese tutoring courses in that time plus easily another 70 hours of home study. I used tutors sourced from iTalki and TakeLessons a few times, but mostly one tutor through LiveLingua. I mostly used the Japanese for Busy People (Kana version) and Genki I textbooks for vocabulary and grammar as well as the Tuttle flashcard sets to help learn Kana and Kanji.
I did end up learning a fair amount of Japanese. I can read all the Kana (both Hiragana and Katakana, about 100 characters total, plus modifications) at about the speed that my 1st grader can read English. I know about 400–500 Japanese words fairly well, including many basic verbs, numbers, colors, directions and questions. I recognize about 100 or so Kanji fairly consistently.
In summary, I can now read through a simple text like the one above, but it’s about at the limit of my Japanese.
The day after my last class, I went on a trip to Japan for roughly 10 days. It was helpful in that I had a few opportunities to use this Japanese for real, actually trying to communicate with people. In the real world, I discovered that I have basic survival Japanese verbal and listening skills.
I successfully asked for directions to a specific train station and understood enough of the answer to find the place. I ordered beers for my friend and I from the train cart lady and understood the amount when she told me what the cost was. I knew when waiters were asking me what I wanted to drink versus if I was ready to order food. I could read important parts of many of the menus. I understood a doorman at a sushi place in Tsukiji fish market ask how large my party was and I answered “two people” correctly. I told someone that I loved ramen when I overheard him asking my colleague if I would be OK with that type of food. I regularly introduced myself to audiences and exchanged cards with reporters. I did all of this in Japanese and all of it seemed to be understood well enough.
I did have a lot of failures though. I couldn’t come up with any phrase to ask where to purchase a SIM card I was holding at an electronics store (I tried something like “これはどこ買えますか”, for those of you how know some Japanese. It drew a very blank stare). I had a few moments with people at meetups where they tried reverting to simple Japanese because they heard I was studying it and I couldn’t understand them at all. I held a menu sideways for a while and couldn’t figure out why I didn’t understand anything until my friend turned it around for me.
Overall, my practical Japanese is much better than my French was when I got to Paris after using Duolingo for three months for 30–60 minutes a day, but that’s a pretty unequal comparison. Overall, I probably spent about twice as much time on my Japanese before this real world test, but at the same time the language is probably four times harder to learn (see below for an explanation of that). So I have no idea how to compare these results, really.
What I Learned
I have a new hypothesis and a new challenge for myself, but first, let’s dive into a few things that I learned during the four months that I tested this theory.
#1. Japanese is really hard
Though in hindsight maybe this should have been obvious, it was not. I severely underestimated the relative difficulty of picking up Japanese compared to French as an English speaker, making the experiment highly unbalanced.
To give you an idea of how much harder it is, the US Foreign Service Institute trains American diplomats and other professionals to become fluent in some 70 foreign languages. They have ranked these languages into several categories based on the amount of time it generally takes native English speakers to reach a level of acceptable fluency for diplomatic work. While French is listed in the first category, taking about 6 months of full time study to become proficient, Japanese is listed in the most difficult group taking nearly 2 years of the same amount of effort.
Not only that, Japanese is specifically singled out in that group as being even more difficult than the others, making it the single most difficult language for English speakers to learn out of all 70 languages.
In practice, two things made progress in Japanese much, much slower than it was in French.
The first is the unbelievable amount of time and work it takes to learn a new writing system. While in French I could at least read the words on day one even if I didn’t understand any of them, in Japanese it was months of constant work to get my reading level up to the speed of a first grader.
I memorized the 100-ish symbols in a few weeks, but going from a 2–3 second per symbol reading rate (which takes like an hour to get through a paragraph) to a more instant recall rate for nearly all the symbols takes months. It makes reading so slow for so long, it’s really difficult to use any text as a meaningful language input source for a really long time. This is exacerbated by the fact that Japanese has no spaces between words, meaning that looking stuff up can be really difficult, since you don’t always know when words begin and end.
The second big thing is that there are so few sounds in Japanese compared to English (around 24 phonemes compared to nearly twice that in English) that it made it difficult to come up with mnemonics for all the words. If you’ve ever studied it you’ll probably agree that there are only so many things that Shoes and Sue and Ray’s Ma can do in your imagination. This makes it really, really hard to form a framework to remember vocabulary effectively as an English speaker.
In the end, Japanese was a poor language to pick in order to test my theory. While I was simply choosing a language I thought would be useful since I travel to Japan occasionally, I was not choosing a good example to compare my French learning model with. The point for me is to see what the best way to pick up a real comfort in using a new language is, but choosing such a vastly more difficult language makes it very hard to compare directly.
#2. Learn Vocabulary Before Grammar
One of the most difficult things about communicating while in Japan was my very limited vocabulary. 400 words may seem like a fair amount, but it’s actually extremely limiting.
I had no consistent review system for learning vocabulary and as I mentioned, Japanese words are very hard to memorize for me. I would simply go through and categorize some section of my hundreds of flashcards I had made and they would get lost or mixed up.
Then I would try learning some grammatical point like comparisons (“I think X is better than Y”) and have nothing much I could slot into X and Y. This made it really difficult to internalize the grammar point because I would never be able to use it effectively.
What I should have done is use a computer-based spaced repetition system to enforce my vocabulary study. I found this system after reading through two books on anecdotal language learning by polyglots. The first was Fluent in 3 Months by Benny Lewis and the other was Fluent Forever by Gabriel Wyner. Both of them cite spaced repetition as the best way to learn vocabulary.
In fact, both of them cite the exact same program to use in order to do this, the open source Anki. The main issue with Anki and the reason that I didn’t use it in the first place is that it has possibly the worst UI of any program I’ve used since I rocked KDE applications on Linux in the late 90's.
It is an extremely slow and aggravating process to add cards to this and you have little insight into your deck.
This is unfortunately one of the last things I learned and still somewhat of a theory, but using Anki I learned a number of phrases in the last month that were very helpful in Japan. It appears to actually be a very effective system.
I believe that a much more efficient way of learning vocabulary in the 4 months of this experiment would have been to skip the manual flashcards and simply taken the time to fill out an Anki deck as new vocabulary came up. I believe that would have made my study time much more effective.
#3. Speaking in the Target Language is Insanely Useful
The first three months of study, my tutor and I were mainly just going through the Genki I textbook section by section. This mainly consisted of studying some grammar point and then doing some exercises so I could repeat it.
Almost none of this was ever used when I actually got to Japan. Some of it was occasionally useful when reading but in real world scenarios it was extremely difficult to pull this knowledge, much of which I forgot due to poor review practices.
Last few weeks I decided to switch our focus to communication scenarios. I had my tutor come up with scenarios and a number of phrases that might be used in those scenarios and then we role played. She would prompt me with something like “You just landed at the airport, ask me where the taxi stand is” and we would do a few minutes of a real conversation. She would throw in random stuff that a person in that scenario might say and I would have to react.
We did this every day, the same scenarios over and over, for two weeks. It was insanely useful. This practice resulted in nearly 100% of the language successes I had in Japan. These scenarios were the only ones in which I was comfortable speaking or understanding a range of spoken Japanese, because I had done them over and over, slightly different every time. I had practiced introducing myself, picking up spoken amounts of money at real world speeds, ordering food off a menu, asking for and understanding random directions.
I feel that had I been studying in this style for three months instead of three weeks, folding in more advanced vocabulary as I learned it, rotating and repeating scenarios over time, I would have been far more effective when I got to Japan.
All of which brings me to the main thing that I’ve learned.
#4. All Online Language Learning is Flawed
I have now used nearly every major online language learning service, many of them very heavily. They’re all flawed in one way or another, forcing people who actually want to gain real language usage skills to combine several methods themselves.
The fully automated ones (Duolingo, Rosetta Stone, etc) are totally useless for listening and speaking skills, which is what I’m mainly interested in. Podcasts are nearly useless for speaking skills. Which leaves us with the online services that have some sort of live chat with native speakers as the only somewhat effective way to gain comfort with speaking and listening in real scenarios.
The main issue with nearly all of these services is that tutors mainly have no idea how to effectively teach a single student. While you can get quality and affordable tutors at iTalki or LiveLingua or TakeLessons, none of these have any built in curriculum and little consistent pedagogy. Most of them seem to rely on uploading images of textbook pages over Skype, which is probably not even technically legal. Even the ones whom I would consider good don’t do much review or spaced recall to make sure the concepts stick or that you’re comfortable with them. Which makes sense because they don’t really have the tools to do so effectively.
In each one, it was mainly up to me to determine what to concentrate on and at what speed, which is insane. Why would I know how best to learn a language?
It appears that in order to really learn a language most effectively, one has to develop and plan your own curriculum and then very specifically tell your hired teachers how to teach it to you.
I would think it’s just me, but both of the polyglot books I mentioned earlier basically say exactly the same thing. They make complicated, usage frequency based Anki decks manually and hire people on iTalki or use a language exchange to practice specifically what they want to. It’s extremely time consuming and sadly a lot of that time is consumed not actually learning the language, but doing the busywork around learning the language.
The only language learning site around that has it’s own curriculum, gives you a sense of progress, lets you test into your current level and uses native speakers seems to be Lingoda. I haven’t used it extensively yet, though it appears to still be up to you to figure out how to study and to do your own review. Furthermore, how to actually book lessons is extremely confusing. I may try using it in the future, though it seems that like the other systems I’ll still have to do a lot of my own prep and review work.
There are also language exchanges, but they’re even worse. It’s very hard to find a partner and rely on them and extremely awkward to actually have the conversations, especially when you’re first starting out.
The bottom line is that there appears to be no online language learning system that you can simply follow and rely on and learn effective second language skills. You have to augment all of them in complicated ways to succeed, which is probably why so few people actually do.
Conclusion and Next Challenge!
In the end, I think that trying to learn one of the hardest possible languages was helpful in testing my theories. I think I learned more Japanese than most people do in 4 months, even if it wasn’t as much as I had hoped. I feel like if I had picked an easier language I would have been fairly happy with my progress, but I’m glad it was harder so I could continue to focus on inefficiencies in the process.
I really want to continue learning Japanese (and French for that matter) and I will come back to them, but in the interest of advancing these theories I’m once again starting over. My family has decided we would like to spend part of our next summer in Germany so I’ll try learning German for the next 5 months, starting now. The comparison will be much closer to what it was like learning French, so I feel like I’ll learn more from the experiment this time.
My new approach to language learning will be as follows:
- Learn vocabulary and phrases ordered by usage frequency using computer-based spaced repetition techniques.
- Do spaced review with native speakers using regular one-on-one video chat sessions using as little English as possible.
- Base all review with native speakers around real world scenarios that I might run into, using real phrases.
- Write a short paragraph or record a short video in the target language every week to practice production and track my progress.
At the end of 5 months, I’ll take a CEFR exam in German to see how far I can get in that time. Wish me luck!
If you’re interested in learning or improving a language, let me know so we can trade notes. Also, if you know any native German speakers who have access to a fast internet connection and would like to improve their English in exchange for helping me with German, drop me a note.