I am occasionally asked how can someone remember 1 mil visualizations. Now 1 mil visualizations is a lot. For comparison, The King James Authorized Bible has 783,137 words. It is enough to know ~5000 words and 5000 phrases of a foreign language to be fluent, e.g. about 100,000 visualizations should suffice per language. Practically, not even a doctor or a lawyer needs more than 200,000 visualizations for their profession. At the same time, it is quite unreasonable to fill any memory structure to its full capacity. Our working memory can usually work with 7 items, yet we start to struggle above 4. So a memory structure comfortably used for 100,000 items will probably be designed for the maximal capacity of 1 mil visualizations. In any case, there is a memory discipline of massive memory structures, and I want to share some insights with you.
Massive memory structures principles
There are some guiding principles designing a massive memory structure.
- Hybrid. Any pure memory structure is relatively small. If you repeat the same element too many times, you will get confused. To get a really large memory structure you need to combine different approaches.
- Large dictionaries. If you need to encode digits, and you use one visualization per digit, it is less effective than using 1 visualization for 8 digits. Your are welcome to check here different methods to encode large dictionaries. If you have 1000 people, 100 actions, and 1000 objects, one visualization can cover 8 digits. However, learning such a dictionary is very hard. It is more practical to have 100 people, 100 actions and 100 objects with one visualization for 6 digits. Even then you may mess up with similarly looking actions. So having 100 people 10 actions and 100 objects may be faster to learn and more accurate to use. In any case, you end up with a very large dictionary with hundreds of objects to learn and apply.
- Revisiting. A huge memory structure contains lots of details you will need to review. If you do not review your memory structures from time to time, you will start forgetting. A huge memory structure typically contains itineraries and clues to guide revisiting process and fill in some missing information if you forget. This information may be backed up by some notetaking system.
- Multiple representation. Some critical pieces of information will be presented in several forms or dual coded for better retention, others will be simply coded for speed.
- Flexibility. As you learn more and fix past mistakes, you may want to add or remove details within your memory structures or restructure them a bit. You do not want this process to be too hard. Some things need to be rigid for speed, yet most of the decisions can be revisited and corrected later on. This also means that a lot of available “storage” space is left empty, so you can add information later on.
Personally, I use a methodology I call “mental cloud” with distributed and connected memory maps similar to the Internet. Several times I tried to explain what I do, and failed to make it sufficiently tangible. So instead I developed two methods which I can easily explain. One is called the mental city, the other mental forest. I will give you the rough structure here. For the details, I wrote but not yet recorded dedicated minicourses.
Mental city
Suppose you need to learn a language or memorize anatomy. This means a lot of attention to details. So memory palace is very convenient.
- PAO We start from PAO to encode markers of three words per visualization of an action figure. This is very convenient since most details are very specific and should usually be described by three words. More details can be added to person and object, while action should be very simple.
- Vaughn cube There is a structure called “Vaughn cube” where the PAO objects I placed along the walls and corners of the room. In my writings, I suggest a slightly different structure, modified and optimized for better revisiting. You can use the Vaughn cube with similar results.
- Mental palace Typically we should have a selection of small and large houses we remember and can use. This is like a dictionary we need to learn before we start memorizing. The specific house is chosen and slightly modified to fit the subject within. A mental palace can encode a couple of university lessons.
- Mind streets Very similar to mindmap. Start with a central square, Divide the city into quarters, then main streets and finally side streets. A side street is roughly equivalent to a university course, a quarter to a university degree. A mental city can effortlessly encode an advanced degree.
Mental forests
When dealing with programming or law our focus is more on structures and less on specific details. A memory forest can easily handle an integrated system of 12 software packages 1 mil lines each.
- Leaves We start with leaves, which are specific structures we need to address. Each leaf is roughly equivalent to a flash card
- Branches, trunk, root The structure of the information to remember is roughly encoded in a mindmap, only this mindmap is a tree and has some useful characteristics of a physical object.
- Toys We can place toys on a christmas tree. This is very useful to allow connectivity between various trees.
- Itineraries In every park there are paths. At the top level of a mental forest, there are diagrams of how different trees interconnect. We can treat it as a mental palace element to facilitate revisiting or as a clever way to encode interfaces between different systems.
Mental landscapes
Clearly, we can mix and match various elements of mental forests and cities, creating complex landscapes. After all, forests have camping sites and villages, and cities have parks.
Most of the encoding is bottom up from a particular piece of data to a mental palace or tree and then placing this mental object within a massive memory structure.
The high-level structures are very flexible to embed these products.
One of the reasons that are still holding me from recording the relevant minicourses: I want to test how these objects can be effectively used in various real situations: programming, languages, medical and legal studies etc. Fortunately, Anna agreed to help and adapt these structures for our 1:1 students that need them.
If you want to generate a massive mental structure for your own needs, 1:1 with Anna is the best opportunity for now. Contact me [email protected] for details
Get 4 Free Sample Chapters of the Key To Study Book
Get access to advanced training, and a selection of free apps to train your reading speed and visual memory
Thank you for the information Dr. Goldentouch. I have one question:
What method would you recommend for studying for multiple choice tests where you have to recognise sentences as true or false instead of “actively recalling information”, and with a vast bibliography full of details?
I ask this because I feel that your method is more directed towards actively recalling short essecial information about an article (as if I was to expose it to other people), and less about being able to recognize 100% of the information on the article through triggers.
I’m not much of an expert but in my opinion it would be a waste of effort for you to build huge memory palaces for that kind of purpose.
If your only purpose is, as you said, to recognize information through trigger words, and not retrieve it in a detailed and ordered fashion (as you would be asked in an oral exam for example) I would focus on creating chains of fast oral/visual associations that can be automatically retrieved upon looking at a word/sentence, thus bypassing the palace or location step. This is way more efficient in terms of speed for a multiple choice test.
An abstract example: “The sky is red” – You know the sentence is false because when studying you created the following association sky-ski-blue skiing jeans-sky is blue
You can use oral or visual associations, just try to link them together and repeat them several times.
I heard about your internet-like memory structures in Jonathan’s podcast and I was fascinated by the concept. I tried to find more clues in your blog archive because I would love to understand it and be able to use something like it. Cross-linking between different fields of knowledge is, to me, the biggest advantage one can have in any domain.
Nevertheless, the city and forest structures seem like tools I can start to work with.
Great! I plan to release a memory mastercourse where I will explain some details of how the internet-like linking can be encoded in memory city and forest. This will take me several months… Stay tuned.
thank you again Dr. Goldentouch!