The Method of Loci: Build Your Memory Palace | Hacker News Reader

The Method of Loci: Build Your Memory Palace

fs.blog

117 pointsfeross62 comments

While Memory Palace descriptions naturally focus on space and spatial reasoning as being the key ingredients to effectively storing memories, I think the binding problem of associating arbitrary items to specific locations, e.g. your front porch or staircase, is actually much more fundamental.

Bower (1972), for example, showed that short-term associative memory could be dramatically improved by imagining a unique scene involving the two items to be associated [1]. E.g. to remember DOG-BICYCLE (so that when you hear DOG you recall BICYCLE or vice versa) you imagine a dog riding a bicycle, or in this article to remember STEAK-PORCH you imagine a cow sitting on your porch. Since these will be very unique to each pair, you won't accidentally end up associating STEAK-BICYCLE or DOG-PORCH during recall.

In this light, the utility of the "palace" is simply that it's an extremely familiar ordered sequence or layout of items (e.g. PORCH, STAIRS, KITCHEN, BEDROOM, etc), not that it is fundamentally a 2D/3D spatial structure. In CS lingo, this sequence/layout would act as a fixed, ordered set of "keys" for recalling key-value pairs that you stored via the imagery process. For instance, if one used another familiar sequence, say the first scenes of your favorite movie you've seen a hundred times, you could creatively put the items in your grocery list into those scenes and then recall them by recalling the scenes in order and remembering what grocery item you put in each one.

Thus, I think there's a sense in which it's not 2D/3D space that's fundamental to better recall, but rather a familiar topologically organized layout of items/locations/scenes that you can use as the keys to bind new items to in the short-term via visualization.

[1] https://www.researchgate.net/publication/232506583_Mental_im...

Right, the locations in a memory palace are just a form of "peg system" [1].

The advantage it has compared with a "link system" [2] is that forgetting one link doesn't mean you lose the rest of the chain. You can also jump to an arbitrary point in your palace.

You're right in saying the use of a physical location is just a convenient but not fundamental part of such a system.

A system that really pushes the idea that the pegs you attach things to can be arbitrary is SemCubed [3], a system where you create a large 2-dimensional grid of "pegs". It also makes use of the Mnemonic Major System that I'm a big fan of.

[1] https://www.ludism.org/mentat/PegSystem [2] https://www.ludism.org/mentat/LinkSystem [3] https://www.ludism.org/mentat/SemCubed

There’s a few movies that I’ve watched so many times I can basically replay them back scene-for-scene on command. If I get struck by something that reminds of some dialog, my brain suddenly starts playing the entire scene, like laying out the context of the quote.

I also figured that the memory palace concept was sort of like an ordered JSON object, so any memorized sequence should theoretically work. I’m now imagining some tool that lets me load up one of my favorite movies, but then also let’s me drop iOS or Snapchat-style stickers on any range of frames. So maybe I need to remember some recipe… At 01:00 I can place a sticker of an image of vegetable oil, then 01:05 is a salt shaker, then 01:10 is some eggs, then at 01:15 is a photo of the finished product. These stickers would be overlaid on the film conspicuously, maybe on top of character’s heads haha

This also brings to mind the possibility of making my entire world a memory palace if I had the above capability in a set of AR glasses. Notes/to-dos/whatevers could be placed and anchored to different places in my home, workplace, neighborhood, etc. They might eventually become interweaved into my mental representation of that location, so that a specific anchored item always springs to mind when I look at my coffee table, for example, whether I have the AR glasses on or not.

You can (and I do) use the fact that it is a 3D space to store further information (such as ordering and relationship between items).

In CS terms, a memory palace is usually used as a linked list while a peg system would be a hash-table but you can turn the memory palace into an array with both O(1) random access and fast in-order iteration.

The idea of the whole system is that remembering locations is something our brain is good at. So this is using our evolved abilities in a wya that benefits us today. Its very different from just memoriaing a list and then pegging items to it.

Those of us with Aphantasia[1] are jealous of people that can visualize anything at all, let alone a memory palace.

Not all of us are "jealous" of course, but a memory palace seems like an incredibly useful technique if you can master it. I wind up just taking a lot of notes and relying on a Reminders app.

[1] https://en.wikipedia.org/wiki/Aphantasia

One thing I haven’t understood with Aphantasia is this. Do people actually see stuff when they imagine things? Some test I’ve seen ask to imagine some object and close your eyes then ask if you see a black wall. Is it actually possible to see anything else?

For me it's like switching an input to receive data from the "mind's eye" rather than my real eyes. My real eyes are just seeing a black wall and I'm still aware of that, but I'm more focused on the imaginary thing that my mind's eye is seeing. It's much more vague and flickery and I can't focus on it much, but I imagine some people are better at it than I am.

It does not feel the same as normal "seeing", so if I were asked that question I would probably say I still see a black wall.

This describes my experience perfectly. I feel like when people discuss aphantasia online, the word "seeing" gets overloaded to both refer to the act of picking up visual stimuli through the optic nerve, as well as a conceptualization action. When I "see" things in my mind's eye, I'm referring exclusively to the latter, and I think most people with normal visualization abilities do the same.

Due to the confusion of language as well as the testimonies of some rare people with incredibly vivid mental images that are on par with actual physical sight, I thought I might have aphantasia for a period of time, but after reading up on it, I feel like aphantasia is a much stronger condition than most people on the internet think it is. "Just seeing black" when you try to visualize something isn't enough - it seems like that's just a sign of being a person with a fairly typical brain who's using the word "see" in its literal sense.

For example, in the star test (https://www.reddit.com/r/Aphantasia/comments/aioyga/simple_a...), if I was a person who understood "see" in the first sense, I'd say that I was a 1. If I instead switch to using the second sense of the word, I'd easily be a 6.

It feels like any test that asks you to rate the vividness of the item you "see" in your mind's eye will fall prey to this problem. I saw a comment on HN where the commenter suggested asking someone to visualize an elephant in profile, and then ask them which way the trunk is facing to see if they're aphantasic or not. I'd also suggest that if you ever imagined a person running beside your car while you were taking a road trip as a child, you also probably don't have aphantasia.

A better experiment, IMO: https://www.reddit.com/r/Aphantasia/comments/cpwimq/ball_on_...

According to that test, I’d have Aphantasia, I guess. I could answer none of the questions immediately, even though I can imagine a table and a ball and somebody pushing it. I guess visually it would be like if you replaced a ball on a picture by the text “ball” and made the table out of “table” labels.

Why should that be better? I've read that and had a myriad of possible visualizations, beginning with some billiard ball on a pool table(which of them?), over some larger ball on a garden table, to tennis balls(which color? they come in neon yellow, orange, pink and even green!), and whatnot else. It only makes sense to abstract all the possibilities out, and ask how the EFFING EFF should I know?!

> Do people actually see stuff when they imagine things?

Apparently so. That's how my wife describes it. She scores an easy 6 on the "star test" - https://www.reddit.com/r/Aphantasia/comments/aioyga/simple_a.... Myself, I normally score 2-3 on it.

I do a 6 too. And I can do all kind of imaginary stuff with that red star.

There seems confusion caused here and in those reddit comments, mostly from two senses of the word "see". The reddit instructions say

> Close your eyes and imagine a red star. Which image do you see? I see 2, but most people see 5 or 6.

It seems some people are (a) reporting what they actually see with their eyes closed, i.e. mostly nothing, a 1. Others are (b) imagining a star and reporting what that looks like in their mind's eye.

I didn't think that aphantasia meant the (a)-type literal seeing, so I just read up on it again, and the descriptions seem to refer to (b)-type imagining/visualization, like imagining a giraffe then being able to describe what you see in your imagination. It doesn't mean you can actually see it, with open or closed eyes.

It seems this test convinced a lot of people they have aphantasia, as it seems most people are a 1. And I suspect most of those who say 6 are talking about what they imagine, not what they physically see. I'm a 1 in sense (a), but in sense (b) a 6, well, I can imagine any kind of red star, like most people, I believe.

Some of the comments report training themselves to produce closed-eye visuals, which I could believe with enough practice. I don't think that has anything to do with aphantasia, though, and I'm in the same box you are.

That was my initial suspicion as well, but I've addressed this distinction with a few people, and come to believe there's a little confusion.

I am 2-3 on (a) and (b). My wife is 6 on (a) and (b). She reports that she does actually see stuff as if it was there. Conversely, I only barely see the things, and at the same time, I can barely "imagine" them in the sense of (b). E.g. my process of "imagining a star and reporting what that looks like" consists of keeping the words, the tokens "red star", repeated by my inner voice on a loop, and then exploring the graph of associations. It's a non-visual process. I don't "see" it, but I can tell you all about it because I know how stars look like and what properties they have.

Like all things, there is apparently a whole spectrum of visualization: Aphantasia (the black wall) and Hyperphantasia (AR-like superimposition of images upon reality).

What I find interesting is that every single one of us is somewhere on that scale, and we don't quite understand the perspective of someone who is on a different place on the scale. Your own question highlights that ("Is it actually possible to see anything else?").

We are all trained from birth to deal with our senses and mental capabilities as they come, barring an injury. What each of us perceives is the de-facto normal. Learning that others don't have senses just like ours is something of an eye-opener.

Like others have mentioned, dream state lets me get past the otherwise ‘black wall’ state, so I can relate to the other, just frustrating that I can’t access it at will.

You don't "see" it but your brain is looking at something. It's like jumping into a 3D universe in your head.

Reading a book for me is basically the same as watching a movie. Every scene plays in 3D in my head. Every single scene. Which is why I don't read horror stuff. Too scary.

If you read to yourself, do you hear the words read by an "internal voice"? Which you know is different from real, "external" sounds, but you're still processing as audio internally?

If so, you're imaging the sense of sound. Visualization is just doing the same thing with sight. You can still tell that "internal sight" is different from real "external" sight.

Also I believe visualization is a skill that can be improved with practice. It's just a form of 2D working memory and it's known that working memory can be improved with practice.

I can imagine a sound that I’m familiar with. But for images i can think about a form of an object, or imagine rotating it, but it is nothing like seeing. Oh well.

> But for images i can think about a form of an object, or imagine rotating it, but it is nothing like seeing.

I expect if you practiced that frequently, and gradually (over days or weeks) tried to add more details or colors etc you would find it improves over time.

But whether that's actually a worthwhile investment of time is another question :-)

We can still edit their memory palace and have fun like that. Put kerosene on the porch and set fire to it. Make the Donald the doorman wearing a mankini etc

Aphantasia alone won't stop you from using memory palaces. If you can remember what you did this morning and where you did those things, you can use a memory palace. It's more about locations and events than visualizing pictures. There are some discussions and examples here:

https://artofmemory.com/wiki/Aphantasia_and_Memory

https://forum.artofmemory.com/tag/aphantasia

Thank you for the link.

I have tried to use memory palaces. The result is that now I have to remember two things rather than one (the memory palace location and the thing I'm trying to remember).

To be fair, I also seem to have Severely Deficient Autobiographical Memory (SDAM) and that will impact personal memory recovery quite a bit. SDAM is common in conjunction with Aphantasia, and perhaps the two combined are conspiring against me in this instance.

Same with me. Full aphantasia and SDAM.

I think I might have aphantasia, but a while back I learned successfully learned how to use a memory palace...

Arabs used the Abjad system for numbers before the introduction of 0-based counting systems. In Abjad each letter of the 28 in the writing system has a number. Like:

a = 1, b = 2, c = 3, ..., j = 10, k = 20, l = 30, ..., t = 100, u = 200, ...

There is no value in the position (like with Roman numbers where a smaller value after a bigger value subtracts it). So "cab" is 6, but "abba" is also 6. Now you can make words and rhymes to remember numbers. Saves you all the mental images of Sinatra kicking superman :)

https://en.wikipedia.org/wiki/Abjad

There's also the major system, which is probably better suited to English.

https://en.wikipedia.org/wiki/Mnemonic_major_system?wprov=sf...

Of all the memory systems I learned from Harry Lorayne's books, this is the one consistent one I use. Not for long strings of numerals, but things like PIN codes and phone numbers. It's well worth it.

I once tried to write a script that would generate the most efficient encoding of an arbitrary length number using the Mnemonic Major System. Definitely an interesting problem!

Ha, do you have that public anywhere? That seems like it would be fun to play with.

Edit: Here's a really well executed online tool that someone else made: https://pinfruit.com/

------

Went digging through my old files, turns out this was almost 10 years ago!.. and looks like I left it unfinished.

I was using nltk to convert cmudict words into MM sequences using phenomes.

Then I was basically brute forcing all the different ways you could slice the given sequences into sub sequences. Then the idea would be to try and identify the "best" sequence.

You'd want to restrict it to using relatively common words, preferably nouns. Really you'd want to avoid any words that are ambiguous or could be easily confused.

Even better would be able to find grammatically correct sentences that fit the sequence.. but I think it's unlikely you'd have any hits.

Anyway.. this is the broken/unfinished code I found in the depths of my hard-drive: https://github.com/johtso/mnemonic_major

A similar method works for what Apple or Google would call "location-based reminders" and get right about one time out of five in practice.

Want to remember something when you get home? Come up with a few words or some such concise thought that will remind you, then picture your front door - just as you'll see it when you do get home - and think those words or that thought. Repeat it to yourself while you imagine unlocking the door or whatever. Do that a few times - it only takes a moment - then forget about it and go back to what you were doing. When you get home, the association thus formed will trigger and remind you.

It only takes a moment, and I find that even with very little practice it works quite well. The method generalizes, too. It doesn't seem affected by aphantasia; I don't really find myself able to form vivid mental images and the "Ganzflicker" thing from the other day had no effect on me past a mild headache, but this technique works fine. Maybe it helps that the visual component is based on experiential memory, but whatever the reason, it works well enough even for me that I can comfortably recommend it.

The thing I've never understood about memory palaces or location reminders is: do you reuse the same locations all the time? If so, how do you mentally distinguish between your current memorized set, and the set you remembered a dozen times ago with the same palace?

I don't know about memory palaces, but the technique I describe involves no memorization; the association isn't that strong. Once it's served its purpose and I no longer need to remember it, I quickly cease to do so, and I haven't noticed problems with overhang from past uses of the same objects.

For the same reason, I wouldn't expect this kind of - I dunno, "mnestic nudge?" - to last very reliably over significant spans of time. I can do it over spans of a few hours to a morning-and-afternoon day, but I don't recall having tried it overnight and I wouldn't want to rely solely on it if I did. I don't really know whether or how well it parallelizes, either; if I've tried that before now, I no longer recall how it went.

What I'm describing really isn't like a "memory palace" at all, even if I mentioned it in a thread about those. That concept claims quite a general and systematic sort of utility, while this is just one among a set of tools that for me also includes a pocket notebook, phone reminders, a diary, and so on. When circumstances preclude immediate access to any of the others, I use this one instead, and also by choice because without at least a bit of practice it might not work when I do need it. (Also because it's really convenient, with less friction even than pen and paper - and besides, my phone can't "remind me of something the next time I pick up my diary".)

The memorized set goes away fairly easily. However, profesional mnemotechnicians who do several demonstrations a day will sometime take the time to 'destroy' the pictures in between demonstrations (setting them mentally on fire for example) in order to reduce the risk of confusion.

If anyone is interested in a book exploring these concepts, Moonwalking with Einstein was written by a journalist originally interested in doing a piece about the world memory championships who wound up becoming the first American to win the world memory championships. The book traces the history of method of loci and the competitions, the potential practical utility (or lack thereof) in history and today, and the interesting personal story of how the author went from wondering "huh, could I memorize 20 playing cards in order?" to memorizing multiple decks at an international competition.

They quote him in the article: Joshua Foer.

Interestingly enough, he went on to co-found Atlas Obscura (https://www.atlasobscura.com/users/josh)

Lynne Kelly's book, "Memory Craft", explores memory palace techniques used from ancient times by different cultures. She is the current Australian Senior Memory Champion. The most intriguing thing for me is how she has turned her walks around her neighbourhood into her memory palace. I think though she now prefers to call them her songlines, mimicking how the aboriginal people of Australia created memory songlines based on their landscape. This has given her a deep attachment to her surrounds. She spoke on Australian radio in 2019: https://www.abc.net.au/radio/programs/conversations/lynne-ke....

Neat! this is basically the app I'm building. A 3D world where you can place files and web browsers on structures you construct. Although it's less about memorization and more for organization ("I put that on the wall past the pink elephant"). And because I'm a nerd, everything gets saved as JSON so it's easy to hack around.

spatial memory is an underappreciated mechanism of our brains, and was one of the founding concepts of MacOS (that you could spatially remember where files are)

Are you me? I’m building a web viewer for rendering polyhedra in 2D or 3D (so that viewers on older hardware can interact with the same data-space as someone wearing the latest VR headset) - media can be framed and text can be flowed within selections of polygons/polyhedra.

For me it’s about positioning data in a visually unique space, so when I’m browsing my file system trying to remember where I saved something, I’ll have the visual cues of form and pattern to help.

I don’t have an updated preview but I describe my reasoning more in a blog with screenshots of a year old 2D prototype.

I’d be very curious to know where your project is at, eventually I’m going to have to port this to Unity so it works outside of web browsers, but I really want to keep the cross-device compatibility, so I’m working on a data storage layer that can serve both (you guessed it, everything is in JSON)

https://coltenj.com/animated-mosaic-experiments/

We are all of us :)

That's awesome! VR has lots of potential but i'm not (and it's not) there yet. Your project sounds really novel. I've equated mine to a 3D database because you can do inner/outer joins, search, filter, and materialize "views" in 3D space.

I'm nearing a demo stage, been at this for a few years - but wow will i need to revisit my code before I expand on future development - it's a lot of code to sling and compared to my day job doing web dev in MVC, leaves a lot to be desired architecturally.

Great job on blogging about your project - i need to get better about that.

Seems we are. I also hacked at something similar - json being yaml in my case - that turns into a weird threejs 'tower'. I used it to 'loci' encode a couple of data sets - mostly the AWS WellAF. https://waf.fudless.xyz and a blog post that goes with it, https://fudless.xyz/misc-memoholics/

This is great stuff - I'm impressed with how far you were able to take things in the browser. I also toyed with a tower concept - one of the things I settled on was what matters is data leading to data; to get to Y, i have to go through X first. So towers are a natural way to do that. I also think you have remained focused on staying true to the method of loci approach, which is admirable.

Additionally, I think having this sharable and on the web opens a lot of possibilities. I hope you keep working on it.

But man, that color change on your blog really messed with me. I have F.lux installed which adds warmth to whites to help with getting to sleep at night or as daylight changes and your site is like a rapid version of that. I'd prefer either not that at all, or at least scroll position based.

Threejs is awesome. Good idea, thanks - will keep the scroll fader in mind for my next blog iteration (when gridsome finally updates). I couldn't make up my mind on background color so I randomized and fade bounced. I sure as heck aint going to win any UX prizes for that ;)

For anyone wondering whether it's too late to train your memory, I'd say memory is one of those traits you can train remarkably late in life (compare versus language acquisition). I met a memory national champion who didn't start training until they were effectively retired and needed something to do.

The mind does not respond to training as well as muscles respond to training. Short term memory , such as digit recall, is is correlated with IQ and not something that can be improved with training. Maybe you can develop tricks for trying to make memorization easier, but like using a level to lift weights, it doesn't mean you are stronger or have better recall.

I'm jealous of my wife. She uses her synesthesia (inborn association between numbers/letters and color) to remember phone numbers and people's names. I think she can remember all the different phone numbers she has had in different countries over the last 50, or so, years. She can certainly remember all the names of people we meet at different gatherings.

At present, she is cataloging her thousands of photos, including writing captions. Occasionally, she grumbles because she can't remember the name of some out-of-the-way location in France or the name of some obscure scholar in Turkmenistan. Me, I'm happy if I can remember to get out of bed in the morning.

She said that when she was child she thought that everyone just saw numbers and letters in color. She didn't realize till about 20 years ago that she had the gift of synesthesia. I think her father and her uncle probably also had synesthesia. Both had great recall. Her uncle was famous for knowing the names and relationships of thousands of people in his district.

I learned about this technique quite some time ago, and it's really interesting but I have a question for any regular users: is it useful for memory requirements beyond rote memorisation? I rarely need to memorise things by rote, but I could still benefit from an improved short-term memory for example.

It's mostly best for rote memorisation imo, though it is really excellent for that.

Mindfulness meditation most likely would be your best bet for improving short term memory, permanently: https://greatergood.berkeley.edu/article/item/can_mindfulnes...

yeah I already both meditate and try to maintain mindfulness on a regular basis; it's helped a lot, but I'd happily incorporate any additional tools I can use!

i tried this method and it useful for remembering passwords/phone #'s but of limited use for studying and understanding languages/math/coding/etc. might be more useful for things like medicine where you have to remember lots of info.

the most effective way i've found to learn and remember is:

1. flashcards with spaced repetition (like anki, etc.) 2. then teaching it to someone (or pretend like you're teaching a class)/implementing it in code 3. solve a challenging problem. this could be trying the harder problems in the textbook and asking for help from the professor or exploring using it in novel ways.

33 years ago I memorized pi to 110 digits while waiting for my then-girlfriend and one of her friends to get dressed. I was sitting in my friend’s car and he had his algebra book in the backseat and that’s how I passed the time.

I’m convinced I could memorize thousands of digits with a little more effort. The only technique I used was remembering chunks of numbers of varying length. The memory palace method sounds difficult to me.

Memory palaces are amazing! A while back I tried memorizing a deck of cards. I spent several days trying to memorize via rote memorization, but it was quite difficult for me to memorize more than ~20 cards. One day I tried memory palaces and immediately was able to memorize 40-50 cards and finally was able to master memorization of 52 cards after a couple of practice rounds.

If you like this stuff, I would strongly recommend Harry Lorayne's books (How To Develop a Super Power Memory or The Memory Book).

I revcently took the learning to learn course on courseraand in the memory segment they recommended trying this site for memory challenge : http://mobile.extremememorychallenge.com/

I got 84% but had made up —-wild— stories about each photo in order to remember it 2 days later

I've had trouble with the sort of example listed. I might imagine pork shoulder on my childhood doorstep, or a pig drinking to beer in a tortilla dress, but I still tend to forget to think about the memory palace in the first place on my way back from work. How do you get better at remembering to remember the right things at the right time?

Some research has been done about this. THe gains, if any, are minimal. Expecting to quickly recall lists of seemingly unrelated words using this method, is wishful thinking. Maybe it can help, but the problem is you still have to remember the associations.

I would be interested in a reference because this goes directly against my experience (I have used this method to recall lists of random words given by a group of people) and the papers I read.

Reply on news.ycombinator.com