Solution to the Language Learning / Memory Palace Problem

Hey Everyone,

So I have most of this info in the video linked below as I have pretty bad tendinitis so I can’t do a lot of typing. Many people on this website may have run into the same problem that I have that the mind palace is great for remembering a translation of a word but can be terrible for actually acquiring the language and understanding words when they are said to us in a sentence. Well, I seem to have found the reason for why this is happening and it all comes down to Krashen’s theory of comprehensible input. He says that it is impossible to learn a language through learning all the words and then learning the grammar and putting these two things together. From what he has read nobody has ever successfully done this without comprehensible input. So I attempted to adapt the mind palace for his theory and I’m getting amazing results. All I’m doing is reading content and instead of just letting my brain create the images it naturally makes when it reads, I’m mind palacing, therefore creating incredibly detailed and fully constructed images in my head. Here’s the video of me talking about this, any criticism or knowledge of others trying this would be v appreciated!

(The video freezes at the start btw, it’ll stop)


Good points that could not be stressed enough.

Your technique is used by me and many others that have grown to understand all that can be put into a palace.

But focus on the target s language in context and minimizing the presence of first language as much as possible.

Also, as I keep on saying to anyone just accumulating words : "

Find ways to use given info out of loci as soon as possible. Ideally in a real life situation. For language s, that’s easy. Find a partner or tutor, italki is good polyglotclub also. And if not possible early enough, speaking to imaginary friends, speaking a mixture of native language/target language to whom ever loves you enough to let you speak to them and find ing it fun."

Good job on your video, thank you! You could give us a couple of your real life experiences with this in writing here would be very appreciated!


Actually what I do is try to never use my original language and only the target language and to learn new words only through context, I’ll make a guess about what the word is and imagine the situation in the same level of detail as when mind palacing (but not in real life locations or locations I plan to revise or remember necessarily) this is actually just a method to transfer words in the target language into images, movements and emotions like in our first language instead of using translation. It’s just attempting to amplify the images in your head when you read and make them more defined and concrete than when you simply read. [quote=“cameri, post:2, topic:41450”]
But focus on the target s language in context and minimizing the presence of first language as much as possible.


I’ve heard the case for and against starting with vocab and I’ve decided to focus on vocab first. I have an Anki file with 5000 spoken words. I’m working through it and also using memory palaces for the hardest 1200 words and so far it’s great.

I have another Anki file with 10,000 spoken sentences but I’m not trying to memorise them. I’m using this deck to attune my ear so I can pick out words and syllables and just have some general exposure. It’s working very well. When I’ve learnt the 5000 words then I’ll dive into speaking scripted sentences and then conversation.

I’ve taken this route because I like it and I have good compliance. It’s improving my confidence and laying a foundation. It seems to me that some people are more narrative oriented and conversation is more suited to them. I’m not one of those people. I’ll get to conversation in my own time.

‘Better the plan you implement than the perfect plan you don’t.’ - Tim Ferriss

(I haven’t watched your video yet but I will get to it today.)


Thanks for sharing this video. I came to know Stephen Krashen through this post, and saw his first video mentioned by you. Now I will also search and see more of his talks. He says logical concepts which make sense to my ears.

I will also try to implement some of his concepts in my Japanese learning.

1 Like

It was nice experience when I started with German short stories and It was good to learn sentence structure and new words too but never tried using mind palace to remember that story can you please elaborate your method with some examples as I am also thinking to get started with this method.

Nice Quote I guess I heard it many times before but can’t remember maybe I could start mnemonics for remembering amazing quotes.

About your method I must say I can resemble you with myself though I am at initial stage, started with German vocab learning and sentences comes next

Nothing below is intended to say “my way is better” – if that were true then there would be no reason for me to be here trying to figure out better ways to succeed at language learning.

I was impatient to hear the method in this video and once it was reveiled it dawned on me that the difference from my method is that “memory palaces” per se don’t get primary focus.

Rather, my Anki (or similar) decks are learned by any technique available and generally for most hard words by simply linking the sounds or sometimes the text to a picture or story then making sure that the first thoughts I typically have when re-encountering the word will be to that story.

Easy words don’t require that last bit – you just say them and link them to something. Hard words require that the story be “engaged” and this may be one of the places that explicit memory palaces can help (or not.)

As to massive input – I am a believer and have used the method for reading fairly successfully (fluent reading in Spanish in 3 months and excellent level in 6.)

However, my use of the Anki (Pauker at that time) deck was to simply “dump the words into my brain” so that I would be able to have that “comprehensible input.”

So I don’t do the flashcard deck to learn the language – the flashcard deck is to get me to 1000 words – to read easy material, 2000-3000 words to start really reading, and about 5000 words to be able to read most things perhaps with some difficulty and certainly with some dictionary support depending on the language and the difficulty of the material.


  • Massive vocabulary to allow massive interesting and comprehensible reading as soon as possible then vocabular becomes less important as a raw input method.


Dr. Krashen is right about so much and yet it is still likely he has the theory incomplete – and I’ve seen an interview with him (and Steve Kaufman a polyglot) where Krashen himself brings up on thing he possibly missed: How polyglots use ‘grammar’ in addition to massive input.

My main problem with reading as the primary (or only initial) massive input is that it is backwards from how we learned our first language.

We first learn to hear, then speak, then both together, then read, then write (in a rudimentary fashion) finally we are taught some explicit grammar to clarify our writing ability and tune our reading distintions.

Therefore: All children learn “grammar” by listening and using speech – and only even much later by writing.

So I have adopted the follow ‘rule’ for learning grammar – though it’s not yet boiled down to a pithy statement:

Don’t study formal grammar beyond my ability to immediatey use it.

Knowing grammar is harmful when you are trying to speak. Using any grammary rule while speaking will almost invarariably block your speech while your conscious mind stops to calculate which grammar rule or rules apply to what you have already said and what you are about to say.

Speaking by using words and then phrases you have learned allows you to be understood whether the grammar is perfect or not in most cases. It’s actually faster to say something poorly in 3 different ways than to stop and work out how to say it perfectly.

Which brings us to “hearing” the language – I am working through this now in French so again, anything offered is tentative based on ways it has NOT worked for me in other languages and ways it seems to be working now for French.

Words in context seldom sound like words in isolation. We’ve all seen this problem.

But worse, when we don’t know a word whild listening it is not like reading – we cannot afford to stop and think it through, even if successful the next sentence has arrived and we are soon hopelessly lost trying to “catch up”.

This is a problem with any type of translation when listening. Taking time to translate stops our ‘hearing’.

With writing it may interupt us but we can resume at the next word when finished with our search (or research.)

Another problem appears that is seldom discussed but is very serious when listening but also affects reading comprehension:

Native listeners don’t hear THE meaning of a word. They hear the meaning of phrases and sentences and even then in context of the larger communication.

If I say “cat”, you have an immediate image or idea which takes precendence and we might say this is the meaning of the word.

But it is really just the primary meaning, or one of the primary meanings, for you.

When I say any of the following the meaning of the word doesn’t change for you it is understood correctly as soon as the phrase or sentence is heard:

  • My daughter Cat is coming to visit.
  • I used to buy Cat-5 network cables from Amazon to improve my network speeds.
  • Some martial arts use the cat-stance as a good method of attack.
  • The construction foreman told the new guy to go start the cat and drive it over to the trench.
  • Just cat the files to join them into one big file.

Even if you have no real knowledge of my family, martial arts, network cabling details, or Caterpillar brand heavy equipment, and no clue of what “cat’ing” a file might be you will hear those “cats” differently.

Of course. So what?

Well if you stop to translate the isolated word “cat” you aren’t just beyind but actually wrong in many cases since you must hear the following words and remember (now consciously) the preceding words to understand which cat – or even to guess which context and decide you don’t know enough about “this cat” to keep up with the flow of information.

We really need to hear the whole phrase or sentence before ‘deciding’ which meaning is right and that’s probably not how our braings operate.

We (probably) hear the phrases and either accept or reject the entire phrase as comprehensible.

We need those Anki words to read ahead of our verbal understanding when learning as adults but even then we want to know a “range of meanings” for most words and not a precise translation to a single maternal language word.

Better to have pictures and stories.

Usually this is how I notice my “hard words” from Anki being transformed through reading or watching TV into meaning and useful words that I might actually speak or write:

The word appears in some context in the TV show or book and there is usually a literal “ah-HAH”.

That word is almost never hard again, and the mnemonics I used to get to that point of recognition are generally no long necessary.

Some of this starting becoming clear when I realized that learning to “speed reading” French was in some ways easier than doing so in my native English:

For French, the rules of pronunciation were not nearly as ingrained so there was actually less of a tendency to mentally say the word as soon as I stopped trying to verbalize.

Teaching speed reading to native speakers requires not just an intellectually understanding but a way to block or prevent our minds from automatically saying the words because the first step from efficient “normal reading” to true speed reading is to stop verbalizing and merely take in the whole phrase or sentence and eventually entire paragraphs or more.

At some point in this process you are asked to read one line forward and the next backwards. (Yes, it works.)

One phrase from my original speed reading course sticks out here: “A willingness to accept out of order input.”

Our minds can do amazing things if given the opportunity to develop the ‘neural networks’ needed.

Speed reading training is to your brain as weight lifting is to your muscles.

In weight lifting you must do something almost incredibly stupid and worthless IN THE MOMENNT. Lift a heavy weight, and put it back down, the repeat many times even though it hurts.

Then increase either the weight or the frequency to make it harder and harder.

Training in many sports works this way but weight lifting is perhaps the purest example.

So what is the weight lifting for hearing?

Some people believe that “shadowing” or “echoing” as David Tolman (YouTube “French Pronouncing Dictionary” channel on YouTube) calls it is the most effective answer.

I am currently adopting this shawoing method to get to the next level of ‘hearing’ French (I hope.)

Glossika Spaced Repetion audio training (SRS) – the old audio version not the current stupid web site marketing effort – is a way to get at it.

For those knowledgable of how spaced repitition flashcards systems (SRS) work in flashcard programs like Anki it might be tempting to think that you already know what SRS means in this new context.

The Glossika audio version is different though the SRS part is probably using a similar underlying theory and mechanism.

It’s basically semi-random syntaces thrown at you so fast that translation from English in total is impossible and one must learn to product the entire sentence, or at least chunks of it at a whole.

It forces us to move from words to phrases.

Currently I’ve finished about 50 days of 1 hour per day (2 lessons) and I it is definitely helping my speaking and probably my listening though other things are certainly causing at least some of my listening improvements.

Oddly, with 5000+ French head words (Anki) and something like 45,000 LingQ ‘words’ (i.e., inflected forms) plus all the cognates between English and French there are almost no new 'vocabulary words" in Glossika for me. I immediately can’t think of a single example in the first 100 lessons.

The ‘vocabulary’ that Glossika is teaching me is that of phrases and grammar without ever explicitly offering a single instruction directly about formal grammar.

It’s simply using the tensens and the gender and other things correctly in increasingly advanced ways and with steadily growing complexity.

It’s also about saying the same thing with different ‘sets of words’ and about changing the meaning of a set of words by adding or removing a word or by changing the verb tense etc.

Essentially it is a vocubulary of ideas in phrases rather than in in single words.

So where does it fall short besides being as painful and perhaps as boring as weight lifting?

(Almost no one is going to do all 312 lessons for 25 minutes each even the few who bought Glossika. Almost no one sticks with weight lifting long enough to become a body builder. Almost no one who takes a speed reading class does the few simple and very hard drills for one hour daily for the 6-10 weeks required to “get it”.)

Glossika’s not teaching how to hear “fast spoken French” only how to produce very clear French at presentation speed.

Hearing real language requires our brains to interpolate a great deal, or to match patterns of sounds that are very difficult to directly map to the underlying words the speaker is thinking and the listening must comprehend.

Native fluent people can make sense out of something that sounds like barely more than rhythm or perhaps hints at otherwise missing words.

They do this without thought – which is actually a requirement for doing it successfully.

In using Dave Tolman’s videos on this subject it has become clear when using an audio editor or analyzer (Audacity, CoolEdit, Praat etc) that some “words” just aren’t really THERE, but the marker is intelligible with practice.

Practice. Painful practice. Ok, at least time consuming and perhaps not that interesting practice. The kind of practice that even the serious student does “once” and then says either that’s useful or perhaps “I don’t get it” and moves on.

The kind of practic that must be repeated dozens of times, and then done again every days for months.

It’s weight lifting for our auditory system.

Maybe there is a better and fast way. Please tell me what it is.

Now, I’ve probably wasted an hour typing this and probably there is no one even interested enough, and who finds the above comprehensible enough, to even read this far.

So, back to drills.


My first long reply sounds like I don’t care about this excellent video and what has gone before in the discussion.

To be clear, I read and listened to every word and agreed with most of it from both the original poster and the answers.

It just got me thinking about where my current study methods are and wanting to share those thoughts.


This is great!. What a good idea!

I followed a similar path, first trying to memorize everything then finding Krashen and the Polyglots conference. You need subconscious access to the language and memorizing doesn’t seem to give that.

My current method is just to focus on massive input + a slow drip of Anki without English (picture <–> word). I also use Anki for pronunciation with mp3s for each word. (I found that learning IPA and practicing pronunciation made it easier to recall words.)

A have a few questions for you @mcnallymemory. :slight_smile:

  1. are you only visualizing the phrases or placing them in palaces which you later review?
  2. Are you organizing your palaces or placing scenes on-the-fly like how Alex Mullen studies. (I think Dominic calls this “memory towns”)
1 Like