When I was studying literature, I made myself a memory palace with 100 rooms and after going through all the rooms 3 times, I started to confuse the information and I wonder how many times I can use a space or can I use a space unlimitedly?
1 Like
I’ve used the same memory palaces many times over the last decade. But:
- I don’t revisit it again for a while (I use another memory palace next time I train/perform); and
- after finishing the recall of data in my memory palace, I don’t think about it every again, so the data fades effectively.
What you describe is a common difficulty, and you can find a lot of information on this forum describing how more experienced memory athletes deal with it.
1 Like
Thank you for your reply Mr. Daniel, but as a professional do you have a method you use for highly abstract words like Latin or Japanese scientific terms (I use a translation app, I hope I made my question clear)
1 Like
Nope I don’t have a method for those because I seldom need short-term memorization of highly abstract words.
2 Likes
Can you use the same memory palace for two different topics e.g. a book on history and a book on finance? Or better to create a brand new memory palace?
1 Like
@ccastell88
A separate palace is probably best for a beginner.
But using the same palace for multiple topics is fine.
To prevent confusion you must leave yourself a breadcrumb trail (my terminology).
Each image must clearly contain references to the images that came before it.
I do this periodically and there’s no mnemonic confusion.
3 Likes
I always revisit the same locations in ml games because I can’t visualize mental images, things rarely merge with the the old games. Instead, all I perceive are shapes of numbers on the screen that vaguely resemble images. I think shape system could improve the experience of using same locations by prioritizing storytelling over visuals
4 Likes
Thank you Daniel 
2 Likes
I experimented a bit, and for me it was 3 to 4 subjects in a memory palace before I started mixing up scenes/mnemonics. Or rather I ended up remember the 3-4 subjects at each station at the same time and it became too much - I would remember plants when I wanted musicians, etc.
It was for fun so I didn’t really try very hard to find ways to differentiate (using different athmospheres for example, which usually gives good results for me).
2 Likes
I also can’t visualise images, and so forget them quickly in the memory palaces I use for competition type training. (I no longer compete but like training on Memory League!) I can reuse my short-term memory palaces with ease.
In my memory palaces for permanent storage - they are all physical and I walk them to encode.
Dominic O’Brien helped me when I started memory training (yes, I do realise how incredibly lucky I was!) and when I was struggling, he said to make logical connections, not just try to visualise something at that location. So I find having multiple things at a location helps me, because I find logical connections with something that is already there, even if it is from a different data set. I never seem to confuse data sets.
Most of my memorisation is for permanent use. Two of my major palaces are for vocabulary - French and Chinese - implemented very differently. Without memory methods, foreign languages were impossible for me.
For English and French, including abstract words, I use Bestiaries. I have a beast (animal) for every combination of the first two letters. I started with the main combinations I needed, and then added a new beast when I had a word for which I didn’t yet have a beast. I never add any now. Some ‘beasts’ are other things, when there wasn’t an animal around with the right first two letters.
I also use a Bestiary for Chinese initials (pinyin) linked to a memory palace based on the radicals of the characters. Finals are done by actions - a PA system. That will make no sense to anyone not familiar with the way Chinese works, but without those systems, Chinese would be totally impossible for me.
I am really into Bestiaries - animals who connect to words. My creatures have become very familiar to me and developed lovely independent personalities over time! And animals can interact with anything, adapt to any verb and even help with really abstract stuff.
Lynne
8 Likes
Thank you very much Mr. @Hari-P you are very helpful for amateur people like me.
1 Like
This is very interesting. My experiment was only with similar data (keywords on different themes for each letter, for a game), but I have other ideas to experiement with and I’m encouraged to know that it is possible.
4 Likes
I wouldn’t try similar data sets. So I am really interested to know if you find a difference.
Let us know how you go!
3 Likes
Hi, Beau, would you give some examples of what you describe as breadcrumb trail? Sounds interesting.
1 Like
Sure @DanteGaxiola,
Each image or scene must clearly contain some link to the images or scenes that preceded it.
My letter frequency palace runs along a similar route as the beginning of my Python documentation palace. They both begin in front of a gas station down the road from me.
The Python documentation begins with Keanu Reeves (as Constantine in the movie Constantine) losing his teeth as False teeth come spitting out of his mouth and onto the concrete.
constants = [False, True, None, NotImplemented, Ellipsis "...", __debug__]
The next scene ( for [True, NotImplemented]) begins at the pumps next to where the teeth land.
Even without the teeth, the clear association of the subject matter
Constantine = Constants
FalseTeeth = False
clarifies that this is a palace of constants for Python code. When the scene for True and None happens at the pumps together, I’m already aware they were preceded by a reference to the categorical type that defines them (“Constants”) and the opposite of True (“False”) since False isn’t included in the True/None scene nor the following NotImplemented/Ellipsis/debug scene.
These are breadcrumbs that help me recall what came before them.
The letter frequency palace begins about the same place (moved over just a smidge to happen at the edge of the street rather than the sidewalk).
letter_frequencies_percentages = {
E: 12.49, T: 9.28, A: 8.04, ...
}
A giant Eagle gets accosted by Tony (a buddy of mine), a few yards down (in front of the pumps) a huge Tornado has swept thousands of balloons up into it, but is being weighed down by an anvil. Just down from that (next to/in front of NotImplemented/Ellipsis/__ debug __ ) a large golden Angel hovers ominously over a snow man that contains Zoro frozen within it (of course, he quickly comes to and slices his way out of the snow man).
Likewise, the letters and their corresponding frequencies are distinct enough and clear enough in their meaning that I can’t confuse, say, Keanu Reeves as Constantine with False teeth for a letter and a percentage.
I’ve left logical breadcrumbs to retrace my steps. No ambiguity remains.
In another instance, I have Brazilian Portuguese verbs in the same palace as my martial arts forms.
I interact with some of the Portuguese images/scenes but not all or even most, frankly.
However, the forms are movements. Like a dance. There is no chance of confusing the beginning kick, block, cover motion of Ki Cho Hyung with my imagery for Acontecer, Aparecer, or Bater. The scenes do interact with each other here (it’s in a former bedroom of mine) but the dance/movement imagery is very distinct from the Portuguese.
I do a similar thing for a palace at my mother’s house. Except it’s Portuguese verbs and Network Engineering concepts/topics.
The submarine on the back patio cueing me to remember “subnets” is logically very distinct from the computer/overflowing cash register of “comprar” in the same location.
I clearly use a lot of “sound alike” mnemonics in these examples. Mainly the logical functionality of imagery leaves the “breadcrumb” trail for me. I know what types of images I’m looking for (networking concepts versus ‘er’ verbs).
Other possibilities exist, I believe.
A breadcrumb is merely my term for a clear indicator of what preceded an image or scene thereby clueing me in to what I should be looking for if I wanted to go backwards in the story/imagery/palace.
Clear as mud? 
4 Likes
Unfortunately, I don’t have a solution that I found myself. @LynneKelly
1 Like
I have found it really depends on what I am trying to achieve. Sometimes I will connect the new piece of information with the original piece of information which I find creates some really creative connections, but I have enough locations so that I could probably encode information for a few months before having to do it.
I’d give yourself some leeway to experiment, but fundamentally in the beginning I think erring on the side of coding things individually is better.
Presumably it also depends on ones personality / mental disposition. If you are extremely good at filtering out extraneous information, it might be easier to maintain a sequential sequence of images from context2, without having context1 interfere, but if you struggle with information filtering you will probably have a lot of ghost images that interfere and make it difficult. This probably correlates with prefrontal function and where one is on the ADHD spectrum.
3 Likes