Making memory palaces scalable

Preface: This is probably only interesting to people who use memory palaces in non competitive settings, e.g. studying.

Memory palace reuse is an attractive idea marred by the scourge that is the ghosting effect.

However, the alternative (more palaces) is so annoying to many people, including myself, that I think some effort into looking for hacks/fixes should be invested. A collective effort, that is. :slight_smile:

If you guys have any ideas, please share them here. I’ll start with the most obvious myself: Making the palace a function of the context.

To me, memory palaces are defined by their loci, the connections between them (the path segments), and their surroundings.

My idea consists of acting on the loci to make them depend on the context. However, this should not change the locus itself, for two reasons:

  1. It is not scalable, for if one decides that they would change the color/shape/size/…, they would soon either run out of ideas if they decide to apply one transformation per locus, or have to deal with a combinatorial explosion of these properties, in which case it will be hard to recall which transformation was associated with a given context
  2. Changing a locus itself implies more mental effort, since we are now trying to recall both the palaces and whatever is encoded within them

My inspiration comes from the fact that many people store multiple things per locus. So, the idea is to encode the context and store it with the other encoded object.

Of course, the context image is constant.

Locus1(Context Encoding, Encoding 1) → Locus2(Context Encoding, Encoding 2) → …

I am looking forward to what you guys will come up with!

2 Likes

Scalability is something I also have been exploring this year.

To keep up with this post and also to contribute a little and synchronize thoughts, I’ll link my forum journal I’ve been using while playing with ideas.

Building Long-Term Mnemonics: Strategy & Techniques (a Beau Journal)

Please do share at least one example of what “acting on a loci” to “make them depend on context” might look like.

Sometimes I confuse myself with all the abstract language we use to play with ideas, but I want to explore your idea more!

Regards,
Beau

1 Like

Many years ago someone on this forum posted the idea to use time as a way to create multiple layers of the same memory palace. So for example you might imagine your garden in the summer and also in the winter. In the winter you might have pineapple trees, winter animals like a snow fox or siberian tiger, a snowman, christmas decorations and in the summer you will see flowers, a swimming pool and whatever thing you associate with the summer. A variation of the winter/summer double layer is the now vs the 80ies. So your current laptop will be a Commodore 64 and your smartphone a cassette player and so on. Once you get into the right flow you can change the whole memory palace into an 80ies version of it without much effort.

I have no idea what the name of that post was, but I thought it was a very good idea.

3 Likes

Thanks for sharing your post!

Say I am encoding information about bats. Then a representation of the context could be a cute bat wearing glasses or something.

I want to remember that bats are mammals, and I am going to store this information in my Kth locus, which for e.g. happens to be a sofa. Then I will see both the cute bat and whatever my image for mammal is witting on it, and potentially interacting with each other.

1 Like

What I dislike about this are the added elements to the palace. I think that it should stay as we see it in real life and devote all of our recall powers to the information that we want to store.

Imagining my room palace is fine, but then trying to make its decor look like something from the 80s requires extra mental effort.

I get your point but I like to share an alternative perspective. I always use my work place as the main chunk of my memory palace because it has much more potential than, for example, my small and boring home. Much of how I visualize my workplace memory palace isn’t even remotely how it is now. The warehouse in my memory palace for example has a setup (assembly line) for a specific customer that we lost more than ten years ago. Some storage places have been transferred into offices, toilets have been installed and removed and so on.

2 Likes

I’m yet to find a good solution for this.

I’m doing a project of memorizing a dictionary and in my first MP I ran into this issue, apparently a classic noob move. I had 25 loci and needed 50. My intuition said: just duplicate it with a different “vibe” and it would be fine, but quickly regretted it because of ghosting. I describe the process here.

I eventually moved the 2 instances/copies side-by-side and created a little story (competitor places, one female-themed, another male-themed), and am currently fine with it, but it came with the cost of reworking both places in significant ways. I don’t struggle with ghosting anymore, but only because I tweaked the ghost-prone areas a lot.

So: the urge to not create another mind palace came with a big maintenance cost, and in my personal experience, it wasn’t worth it.

I suspect a better solution for expansion would be spacial and not qualitative, as most people suggest.

For example, I plan to experiment with mirroring the mind-palace in some axis. But that doesn’t sound much easier brainpower-wise.

I suspect duplications/instantiations fundamentally don’t work well because of some neurological reason (the same reason humans are good at assigning space to memory contradicts the idea of reusing the same place as a different concept), but I’m curious to know if someone more experienced managed to hack the issue.

One caveat is that I have bad visualization skills, not exactly aphantasia, but it’s almost like I can render a scene in my brain but the monitor is off. I know it’s there, I could draw it, I could vividly dream about it. But adding vibes (“winter”/“summer”/“in flames”/“frozen”) doesn’t work well for me.

4 Likes

This is something I do. The bat is an avatar, charged with embodying the topic you are encoding information about. In one scene your bat might be screaming at a dolphin, which in turn is screaming back. Why? They’re arguing over who has the better echo location.
Another scene it could be sitting on a woolly mammoth, so you can get mammal out of it. Maybe the bat is teasingly calling it a woolly mammal.

I don’t yet know if an avatar (character, rapscallion, hero, whatever) work as context to to re-use loci, but it seems plausible. Instead of just looking to see what is on the sofa, you’re checking to see what’s there relating to Prof. Night.

2 Likes