Study: Mnemotechnics in Second Language Learning

Interesting study about using a keyword association method when learning foreign language vocabulary:
Mnemotechnics in Second Language Learning [PDF]

88% recall with keyword mnemonics vs. 28% recall for those who didn’t use the technique.

See also: Emerging Technologies from Memory Palaces to Spacing Algorithms: Approaches to Second-language Vocabulary Learning.


That is a huge difference!

1 Like

Great article! Thanks!

One thing that interests me is that students sometimes have to go back and add detail in order to make the memorization “work.” I think many people hesitate to do this when they are testing. I know that I have.

But the truth is that there is nothing wrong with making the association in the location stronger. Yet, people whom I have tried to teach memory techniques to tend to think that if it doesn’t work and hold the first time, it’s game over and the techniques don’t work. It’s frustrating.


NVM, this is interesting. Thanks for sharing:)

Great paper Josh,

I really like these scientific papers that have actually done some experiments to see how a specific technique can affect our ability to learn. I have found another one that actually lessens the wonders of the keyword method. However, in both papers the words used in the experiments are always concrete nouns for which it is easier to create the “imagery link”. Verbs are probably one of the most important types of words but yet they seem to be left out and likewise for adjectives, adverbes…

Thanks for this article on limitations, pol.

The first thing I notice about it is that they are using test conditions. I think that most of us would agree that decent relaxation is needed for strong memorization to take place, so I wonder how the results might be improved by a meditation session or some method of relaxing the mind.

Second, a lot of people disagree with me that keyword work should be combined with loci. I’m not sure why, but most people prefer to leave the keywords floating in space.

The personal reason I found locations in combination with keywords work so well for me involve the ability to rehearse the words I’ve memorized in a structured manner, and - I’m speculating wildly here - but I think that having things put in dedicated places helps the mind get over the fear that anything will be lost. We don’t like to lose things, so I think we seize up whenever we have any anxiety whatsoever that something important to us might go missing.

As ever, there needs to be focus on exaggerated action in the mnemonics.

Thanks again.


metivier is very true what you said
I think it is always essential to combine the keywords with the method of loci

1 Like

Metivier, I totally agree that locations should be used. I personally haven’t done so (!!!) but it’s so obvious to me that I should be.

Doing duolingo italian, I was able to get through a majority of the lessons on my first try, simply by using mnemonics. And my recall was quite good. But, for years I’ve been wanting to create enough loci to place all my new words in each day. Ideally I would have one journey for each day of the week with, say, 100 loci. I would load up new vocabulary each day, and review the previous days. By the time next week rolls around you should have those other vocabulary pretty well fixed if you have reviewed them.

This isn’t even touching on using towns for each language, but as I said, I haven’t put in the time to make enough journeys.

I think locations is one option, but I’ve never found it to be necessary. And I do somewhat question whether having the number of real locations that Metivier suggests is really any more useful than other methods. That’s not to suggest that it’s a bad idea, just that it’s not the only one.

Personally, I tend to tie my mnemonics into memories that I already have and I don’t even bother writing down what the mnemonic I used was. The main reason being that I’ve got motivational problems most of the time and in practice coming up with a new mnemonic every time tends to wind me up in the 85-95% range most of the time anyways.

Locations simply give you another chance to recall the images and when I memorize stuff, I want triangulation (location, imagery and meaning). There are other ways to amplify this as well using Memory Palace journeys.

In addition, a location-based memorization approach is most likely the only approach that lets you use other memory techniques within itself. I suppose you could have a Memory Palace within a link or chain, but I’d rather have a link or chain of Memory Palaces, each of which are loaded with links and chains. Makes much more sense to me.

I always teach location-based strategies because it’s one of the easiest ways for initiates to experience a quick victory. Yes, different people can do all kinds of different things with their mind when it comes to memorization, but we have an ethical duty to share these wonders and one of the best ways for a person to “get it” in a way that is quick, easy, efficient and elegant is to teach a location-based strategy.

Finally, not to get semantic or philosophical, but all memories are “located.” There’s no getting around it. As pfrank says, he “tends to tie” mnemonics together, which draws our attention to the primacy of location in many ways when it comes to memory. Even to conceive of memory requires us to think of the “place” where it is “stored.”

On top of that, all memories are created by a mind as it rests (or works or plays) in a body that is in a location. It’s well-known that the brain secrets noripinephrine when we find ourselves in unique and novel situations, which is why we tend to have very profound memories of “the first time” we visit a place or experience a thing.

We can readily capitalize upon this chemical behavior by creating a multi-layered memory experience:

Develop and use location-based memorization techniques (i.e. carefully constructed Memory Palaces) while in a unique and novel place.

As is happens, I tested this during a recent visit to Athens …

It worked gangbusters. All my Memory Palaces created during the trip were (and remain) super-charged and the Greek vocabulary I picked up massively intensified using my location-based approach, known to some as The Magnetic Memory Method. It works pretty damn good without all the trouble of going to Athens, but the point about noripinephrine remains the same. Used with caution, it’s fantastic stuff for language learning in combination with a location-based memory strategy.

1 Like

I’m sorry I didn’t get back to you sooner, I got busy.

Anyways, I think memory palaces are perfectly OK, they clearly work quite well for people and that really should be the basis upon which techniqes are evaluated. It doesn’t matter how well a technique works for somebody else, if it’s not doing the job for you after all.

I think the main weakness with loci is that since you’re associating the material with something that’s unrelated, it’s not quite as strong as it could be. Obviously, it’s strong enough, but I tend to find that I have to do more maintenance when I try to store information like that.

I tend to view memories as being associative as in they’re connected to each other. Trying to apply the idea of location too literally just makes things rather confusing IMHO. There are similarities to the way I place my memories into my brain with the method of loci, but there is no particular order to it, I can traverse my memories in many different ways, so it does nothing good for memorizing things in order. If I need order, then I’m sure to use loci, city files or some method that’s inherently ordered. Most of the time though I don’t care about order.

But, the key thing here is that “placing things into the void” is bad, really, really bad. A memory without ample connections to other memories is unreliable and unstable. Whether one chooses to embed the word into a sentence, a visualization, a song or something else is of less importance than the fact that it’s been connected to other things. And I think that’s why your ideas of bridging characters and compounding are so important in any situation where you’re wanting to retain information for the long term. When I recommend your books, it’s usually not for the loci, it’s usually for those two bits of information because they’re that useful.


I’m 100% agreed on avoiding the void (which reminds of the great Georges Perec novel “A Void” in which he manages to avoid the letter ‘e’ throughout in both French and its English translation!)

I think the relevancy of location can have relevancy and power, but ultimately it’s mostly irrelevant where the Memory Palaces are and any confusion comes down to poor management and “pilot error.”

I’m enthused to hear that you found the compounding and bridging useful. In my experience interacting with readers, virtually no one makes use of these core principles, which suggests to me that in future editions, these need to share more of the stage with the location training stuff. Thanks for making that observation.

1 Like

I’m a bit surprised by that because it’s a lot less work and quite frankly I’m too lazy to roll out the full method unless I have to. But, the bridging and compounding techniques take very little investment ahead of time and make a huge difference. The bridges take a bit of thought ahead of time, but not anywhere near the amount that it takes to set up a palace for every letter. Heck, you don’t even have to create the bridging character ahead of time, even if it’s better, you can also do it when you encoutner a number of related words.

Then again, I’m studying primarily Chinese and memory palaces for that require more fiddling than what you would use for most other languages. If I wanted to memorize the dictionary like people do for other languages, I would have to have 214 or so palaces for the radicals, an index of some sort for the stroke number and then I would need to store the word. Needless to say, sometimes trying to stick too literally is a huge mistake.

Yes, wonderful to see some stats on it! Thanks for the information!

Yes, being too literal is a huge mistake that people make. That’s why I prefer the word “method” to “system.” Systems tend to me fixed and rigid. Methods are flexible.

I’ve recently discovered a way that you can get 214 radicals into a single Memory Palace that is so simple and elegant. I can’t believe I didn’t think of it before. In fact, it’s a way to turn a single Memory Palace into 520 Memory Palaces in about 5 minutes of your time.

Feel free to get in touch if you’re interested!


Great material for language learning. As a student of many Oriental languages, any method or system is quite helpful and useful in successful mastery.

Hi man , yes i’ am so interested ! How do you do that? I Have bought your book …Thanks for all these posts on the site , it 's very helpful