I know creating connections between consecutive loci can improve their recall later on. However, I was thinking of using this connection to encode information as well.
For example, imagine you’re using PAO; actually, let’s imagine you’re using something a little bit different, let’s call it PAP: a person making an action to another person (instead of a person making an action using an object, like in standard PAO). With PAP, you could imagine the second person of the first locus making another action to the first person of the second locus. Using this connecting action to encode digits too would make the memory palace both more memorable and more efficient. The result would be a continuous memory palace without clear division between the loci:
PAP-A-PAP-A-PAP-A…
What do you think about this idea? Have you tried it? Do you expect any drawbacks from using it?
Sounds interesting. I’m looking to develop a continuous system too so that I don’t run out of loci for unending things I want to memorize e.g. proverbs, quotes, jokes.
For example, imagine you want to memorize the number 913059941926094612742. “Batman slapping Bart Simpson” could be the first locus while the next could be “Superman shooting Grinch.” In the middle, we could imagine an action like “punching.”
Ok, I think I get what you mean. After I finish my 3-digit PEG list, I’m going to test this memorization method and see if it works well. I hope having 1000 different people and 1000 different actions will make it differentiated enough to avoid ghosting. Thanks for the advice!
Thank you! I’m struggling with 1000 distinct people already
However, I got an idea for finding actions. Since I’m going to put people acting on other people, instead of people acting with an object, I can use names of objects to remember actions. This can substantially increase the number of actions. For example, “taser” can be seen as the action of using a taser, so action 104 is checked . This reduces the size of the peg list from 3330 to 2220 (3-, 2-, and 1- digit lists) and also makes it easier to come up with actions. I did this when I created a 2-digit peg list. Do you think it can work with a 3-digit peg list too?
Thats not a bad approach. Its VERY difficult to find enough distinct “object independent” actions to fill a 1000 list, especially if each entry also has to follow some kind of rule whether thats categorical, phonetic, or stemming from a person.
It might be easier (and give your scenes more consistency) if you just did 1000 objects…
Everyone is different obviously, but I’ve found that its easier to have the action not matter vs. having the object not matter when constructing scenes.
Have you thought about trying for a 1000 person + 1000 object list and either letting the action flow intuitively to connect them, or finding 100 distinct actions for a 2-digit set, so it becomes 3-digit person and 3 digit object connected by a 2-digit action?
Hmm, that sounds like a great idea! I haven’t thought of that. When I created my 2-digit peg list some years ago, I used to personify objects since I didn’t know many celebrities/people/cartoon characters. Now this isn’t a concern anymore, but I was going to do the same thing to the 3-digit peg list to save time. Your idea seems much better, though.
I want to have three independent lists for 3, 2, and 1 digits, so this wouldn’t work.
Having a thousand objects sounds much more feasible. Verbs are indeed difficult to distinguish, even more with this idea of continous memory palaces. Since each locus will be connected to the next, verbs like “swallow” will be hard to visualize because they “destroy” the person who is receiving the action, so I wouldn’t be capable of imagining them acting on the next locus. I was thinking about this problem yesterday, but your solution coincidentally solves this too! Fortunately, I haven’t started filling the action columns in my excel spreadsheet, so I won’t lose any work by trying it.
Have you done this in practice btw? How’s your peg list?
I just completed the 1000 objects and 1000 people lists. My original/current 3-digit list is a single element per number, a mix of people and objects. It was built for the Shadow System for playing cards which only requires one image per number. I’m fairly fluent with it, still have more to go to meet my self-imposed goal. But along the way I started filling in options for both categories because I eventually want to get to a full 2704 image 2-card system and a 3x2x3 PAO system.
I’m focused on a goal with cards at the moment, but once I get there, I’ll move to start developing fluency with the new half of the PAO. My previous experience with a 2-digit PA(A)O (adding an Adjective category to give me 8 digits per scene) was really effective, and got me to 4325 digits of Pi memorized very early on in my exploration of mnemonics. One thing that I’m happy/proud of myself for doing was being picky and thinking future-proof when evaluating and filling in my 1 and 2 digit lists to make the upgrade to a 3-digit and card system as seamless as possible.
With all that said, I’m still VERY new to all of this too and far from a pro, so take all of this with that in mind! Try or at least read and think about a few different approaches and if one seems to click and make sense to how your brain likes to organize and process things, start there and see how it goes. When moving into a 3-digit system, the best advice I can give is to try to be as consistent as possible with the rules for construction that you choose. Try to allow as few exceptions to those rules as possible. 1000 images to learn and develop speed with is a VERY big commitment in both time and mental energy. Take that effort up front in the build phase to save it in the learning and application phase.
(Here’s my full 2-digit list, including people, actions, objects, adjectives, locations, and cards. It is all based pretty strictly on major system phonetics, except a few of the locations. I haven’t posted my 3-digit list(s) yet, but happy to share some examples if you have any specific questions.)
There’s nothing wrong with using 2-digit actions to link your 3-digit people and objects. It lets you compress 8 digits into a single three-element scene. There’s nothing that says you can’t combine lists. If you have a 7 digit sequence, you can’t really use just 3-digit images, or just 2-digit images. You need to use 3 elements minimum, whether its structured like 2-2-3 or 3-1-3 or whatever combo, you still need to dip into different lists to encode it.
Look at an 8 digit sequence, something like 602-11-104 could be a simple self-assembling three-element PAO scene: “JaSoN TooTing on a TaSeR.” The same sequence could be split up using elements from all three lists, 60-21-1-104 could be “JuiCy AnDy eaTing a TaSeR.” More elements, so not as effective, but maybe this scene is more memorable to you. Having the option to transpose is kinda nice. Its not that you’re duplicating lists, you’d use them contextually when they fit. Thats the advantage and versatility of combining lists to give yourself unique options for scene construction.
I understand what you mean. Indeed, you can combine the lists, and I do intend to do that in the future for memorization of numbers. However, many of the things I want to memorize are already chunked, and they often have 3 digits, so I need to have a 3-digit peg list that can work alone too.
Oh. Advise against a continuous memory palace? Or linking the way OP said?
I don’t like to use small palaces eg houses because it gets crowded in my mind. So i use memory journeys. With the typical overenthusiasm of a beginner, I’ve memorized more than I thought I could, and now I’ve started running out of routes in familiar with
I’ve seen many posts to come up with new journeys, but I was wondering if there’s a continuous journey available, why not!
I was thinking more about your idea. When you said to have 1000 people and 1000 objects connected with 2-digit action, did you mean that there would be a person at the end too, like a person-action-object-person (person acting with an object over another person)?
I was thinking more about this, and maybe it could work quite well for what I need. With this idea of continuous memory palaces, each locus would have:
Like a classic PAO structure. Person from 3-digit list does Action from 2-digit list to/with Object from 3-digit list. 8 digits per scene.
Any particular reason you’re trying to combine so many elements at a single loci? I’ve found it much easier to recall two scenes that each encode 8 digits than one single scene that encodes 16.
At one point I was working on memorizing binary sequences and I experimented with a 9 element scene that encoded 54 binary digits, in the form of “Adjective/Person1 does Action to/with Adjective/Person2’s Adjective/Object which then does Action to/with Person3.” It “worked,” but it was really unwieldy to work with and construct. That same 54-digit sequence split across three much simpler PAO scenes at consecutive loci was so much easier to encode and recall.
Remember, the big key is how many digits can be compressed into a single ELEMENT in your system, then how your elements combine to create an image or scene that is easiest to recall. You can combine tons of elements into a scene to give you a ton of encoded numbers, but you still have to encode and decode each element individually. If the goal is to reduce the number of loci, you have to either compress more digits into each element, or cram more elements into each scene so you dont need as many loci.
In my opinion, its much easier to just use more loci rather than try to create more and more complex compound scenes.
You could definitely use an action as encodable element that links your scenes across loci, nothing wrong with that approach, but I’d try to keep the elements per scene at each loci as uncluttered as possible and avoid repeating element types in an individual loci’s scene.
So, Person-Action-Object/(action)/Person-Action-Object could be pretty effective. This way you’d also have a natural prompt that after the “Object’s Action” is when you move to the next loci. It’s easy to lose your place if you have a chain of elements that contain a repeating pattern within the same scene. “Was that Person I just encoded/decoded the second person or the third in the scene? Is it time to move to the next loci or is there another set of actions and people at this one?” By simplifiying the structure and avoiding repeat element types, this confusion potential is pretty much a non-issue.
I neither compete nor intend to memorize very fast. My goal is to store factual information from textbooks or articles: tables, lists, specific information, historical dates, research papers’ results, etc.
For instance, in the case of the periodic table, the 1st person would encode the name, the second would encode the symbol, the objects would encode values of the atom that may have up to 3 digits (like atomic number and atomic mass), and the actions would encode values that may have up to 2 digits (like the numbers of the line and column).
If I were competing, I could forget the info just after the competition, but my memory palaces are intended to be held forever. Given this, memorizing two 8-digit loci may be harder than one 16-digit loci since the former will imply having twice as many memory palaces.
It will be harder to get used to it, but I believe anything practiced for long enough becomes fast enough (although it is probably gonna take much longer to get used to 16-digit loci than two 8-digit ones).
Does my particular use of the method change your mind on the issue? Or do you think it isn’t good for that either?
I think there may have been a misunderstanding. I won’t need to think of moving from one locus to another. The last person from one locus will move towards the first person of the next locus, so I will be just following them.
Ah! That makes total sense. I thought you were structuring this for raw digit memorization. I actually really like this approach for how you plan on using it.
I did a similar thing a while back when memorizing the table of the elements, with stats about the planets in the solar system, and with a list of Superbowl numbers, years, teams, and scores.
I think when working with loci and representative imagery this way there is less of an importance on number of loci and more linking within each loci to get a complete set of info, like you were mentioning with the table of elements. In that case, it makes sense for you to use one loci like you’re laying out, one for each element. That sets the stage for the scene you construct to inform each element. Really smart.
I think as long as you’re consistent with which imagery represents which type of info, it will be very effective.
For my periodic table project, I used a main person to represent the number of the element, a representational “soundalike” image for the name of the element, 3-digit word for the P/G coordinate, and a “dialogue phrase” for the atomic weight.
So for Element 5, Boron, row 2 + column 13, atomic weight - 10.810… my image lived in my fifth loci, the main character was han SoLo (5 in Major System), who was dressed as the red Baron (soundalike for Boron), he was reading an aNaToMy book (position 2.13) and he asks if I like his aDheSiVe DiCe (10810 in major) as he shows a big pair of fuzzy dice stuck to his plane.
This is a way more involved set of images with way more total digits than I’d ever use in pure number memorization, but it works because of the nature of the material and because I organized the structure of my data ahead of time.
Something to consider is that you likely will need to cater your scene structure to the info set you’re working with and each topic or set will likely require different structures. So your P-A-P-A-P-A etc structure probably won’t be a great fit for all (or even most) data sets you’re looking to memorize.
The scenes within my palace that holds the solar system info is not organized the same way as the table of the elements, and the palace/scenes for my Super Bowl winners/losers/score project is constructed differently than both.
In this way, you’ll probably need to be kind of organic and approach each project as its own entity. There won’t be a one-size fits all template because different topics will require different types and amounts of data per loci or different approaches to building your scene.
The cool thing is that if you aren’t going for speed and have the time to consider how you want to build your scenes, you can make them really strong. The more time and the more ways you are active in working with the information, the better you’ll understand it and the better you’ll be able to organize it for memorizing.
There really is no substitute for understanding when it comes to memorizing. Yes you could simply convert data into mnemonic imagery and then plug those images into loci with no understanding and you’d be able to recreate tables and pull stats out of your brain with relative ease, but having an understanding first improves the process of memorizing immensely.
So yes, long story short, your approach should work for long-term non-time-sensitive projects! Just be aware that you probably won’t be able to pre-set a one-size-fits-all structure and will need to be flexible with your imagery and scene configurations!