One of the difficulties in teaching yourself a second language is that you need to come up with a method that you can follow and that would allow you to achieve your goal. Not many people can actually do that, which is why, in my opinion, AJATT is so popular. Easily accessible and understandable, AJATT lets those in doubt hang on something and execute.

But it is one thing to have a method and the other to know what can go wrong with a given approach.

Unfortunately “what doesn’t work” can be assessed only after some time, which is more like year/years than weeks/mouths. I would like to create a place (this thread) where people would be willing to share their experience with things that didn’t quite work for them.

Note: Although I did mention AJATT, it is not my intention to create anti-AJATT thread. I do believe that there are people who thanks to AJATT achieved their goals.

For the start, I collected two testimonies I've seen in recent days.
1. The "What's this word/phrase?" thread

HonyakuJoshu Wrote:It is the biggest mistake I ever made - I was awesome at French and Spanish Grammar at Sixth form and neglected Japanese grammar I swallowed the lie it was easy compared to romance languages. I am currently searching for someone to teach me... I think if I spent more time in Japan (I've only spent ten days there) I would be better at grammar.

2. Writing While Reviewing

eggcluck Wrote:I never used to wrtie my reviews I though I was doign pretty good with my SRS ( sometimes I used the finger trace thing, I count that as non writing)

Then one day I decided to write them out...my % dropped to 50% turns out Kanji I thought I had down were in some way not quite right ( in small ways that made big differences) even those I traced out with my fingers that I thought I was getting right so now during reviews I write them all out. Putting ink to paper is a good way putting down in hard form what you actually do know rather that what your brain thinks it knows based of some hazy memory.

More testimonials here.

Yeah, theres already a thread for this.

I think my biggest pitfalls when starting out were:
- searching for and using 'perfect' materials (as a follow up for RTK), things like pre-made SRS sentence decks, iknow, etc.
- neglecting listening
- relying on a shit cellphone dictionary (Get an electronic one)

@nadiatims
So how did you get over "searching for and using 'perfect materials"?

basically I started slacking off on that kind of study, because it was boring, and I was spending too much time trying to find the perfect set of cards to do next, and I realized that I was still learning a lot of vocabulary incidentally completely independent of SRS, etc. I also found there's huge amounts of vocabulary that's used all the time in reality and authentic media that isn't covered in the different graded wordlists. The wordlists are often quite arbitrary selections of words that someone has decided you are supposed to know, and are often skewed towards news and business. It's not clear what the selection criterion is. I also realized that learning words in isolation (even with an example sentence) doesn't really reinforce the word in a particularly meaningful way and it doesn't train any of the other necessary soft skills needed for comprehension, ie. the ability to interpret meaning and read between the lines based on context, guess unknown words and readings based on your existing knowledge and context etc. I realized words are truly learned when they are heard and understood in many different contexts, so it's not worth spending to much time on the initial exposure by processing them through an SRS deck or whatever, and making sure you have audio or example sentences or whatever else. So rather than nailing things down via perfect materials that focus on a limited set of data, I found I learned more (and more useful things) by exposing myself to the language in its whole form.

[Zgarbas]

On the other hand, though I did not spend much time finding the perfect list(pre-made ones are plentiful both on anki and in former smart.fm), I rarely don't recognize a word if i've SRSed it. Maybe if it's a homonym or what not. Core 6k has been very good to me. On the other hand I ran across a list I made previously with unknown words (words that I ran across in native material and studied for a bit) and my retention rate was... irrelevant.

"Don't waste lots of time focusing on ABOUT STUDY and focus on STUDY" is important(and spending lots of time making/finding perfect lists might fit here as well), but the SRS thing depends on the person, me thinks.

Certainly, if you don't waste too much time in the making/finding of materials, and you enjoy it enough to actually use while still leaving time to use/reinforce and learn new things in the wild, that's great. My main point though is that the first exposure while obviously necessary (hey you gotta learn it somewhere), is in the long run much less important than the subsequent exposure and reinforcement in different contexts, so it doesn't matter if your understanding is vague or incomplete (or even wrong in some cases) or pronunciation imperfect at first, so I think when people spend a lot of time searching for wordlists with pitch accents, example sentences, finding audio for words, processing some textbook or media into a deck etc, it really seams to me like more bother than its worth. Why spend so much time on 1 word or phrase when there's another 20,000 or so you really ought to know.