Never break words into letters unless you study Klingon

It is useful to leverage the available information in large chunks. The larger the chunks, the faster and more effective the process. The smallest chunks to remember are letters. The larger chunks may include the keywords that allow access to the entire resource.

The most unfriendly language

Klingon is a constructed language, specifically intended to sound unlike any human language. I can honestly say that I cannot remember Klingon without learning specifically each diphthong. Even in Klingon, some combinations of letters are more useful, and these syllables and diphthongs are larger than specific letters.

If you are not learning Klingon, and most people do not speak Klingon because it is fundamentally complex and useless, you can use larger chunks. We are taught to do this in schools, where we disassemble words into prefixes, roots, suffixes, and whatever else is required. The prefixes and the suffixes will usually repeat, and we can reuse some visual dictionaries for them.  The roots may require some etymology research.

Surnames and names of the places follow the same logic in a different language.  For example “Lagrange” with some effort can mean “that grainary”: pointing finger and a silos visualization. “Newton” can be encoded by a neut, an apple that fell on the scientist, or the famous portrait of the scientist because he was famous.

You can and should break words into their parts in other languages. If you remember “bat’leth etymology” as “sword of honor” and “mek’leth etymology” as “bridge sword”, even Klingon is easier to learn. I repeat, there is no return on investment in learning Klingon and it is here as an example.

The language of details

Quite often you do not even need to remember a whole object: address the most noticeable detail. Take for example the US declaration of independence. For me, the whole document is characterized by the huge John Hancock signature. The American war of independence is visualized by Yankee Doodle. It is fun to remember things from fragments: life looks full of wonder, like it does for a child.

Specific details also enable disambiguation between two very similar objects.  For example, flintlock and matchlock are very similar muskets. Simply one is fired with a spark from a flint, and the other uses a burning wire to activate the charge. The entire musket is just a zoom-out of the simple mechanism used for firing in different ages.

Clearly, sometimes long visualizations are needed for one word. I still struggle with some parts of Hippopotomonstrosesquippedaliophobia which presents a fear of long words like parts of monsters. You may encounter many words of similar complexity in organic chemistry, but hardly anywhere else.

One image worth 1000 words

If you have a description of “continuous increasing infinitely differentiable function”, it is easier to visualize an instance of such function in its  “S”-shaped glory than it is to visualize the long definition. If you have an elephant -visualize the elephant and not “a heavy plant-eating mammal with a prehensile trunk, long curved ivory tusks, and large ears, native to Africa and southern Asia. ” The meaning is the same, so why not use it? And if by some miracle the author describes mastodon, visualize a saber-tooth tiger on in.

Usually, we are not supposed to learn very long descriptions. Instead, we are supposed to generate one very detailed image and remember it. Maybe the author wanted us to visualize, and maybe he did not have enough money to buy a good illustration – the reasons may vary.

If you have a symbol, use it

Quite often I am asked to describe logical markers by one student, and on the same day get a very long and colorful visualization of “+” or “=” from another.  It is not funny. Mathematical language compresses a lot of philosophical depth in very simple symbols. Use them whenever it is applicable. Use simple geometric figures whenever you can visualize them.

Another common challenge is accuracy. The accuracy of pi is limited. You can learn to memorize thousands of digits of the constant, but you are seriously advised against it. Two or three digits will suffice for almost any need. If you need more, use a computer. Even the computer will have issues above 64-bit accuracy, which is approximately 20 digits. For something longer consider CGBN. And again, you are highly unlikely to need it ever. Even computers use very limited precision in tensor processing units. Two digits or four bits are just enough for most use cases.

Visual simplification

Mathematical symbols are very simple. Very complex ideas are often expressed by one or two lines. To make these lines special, we combine them with the context. In math, the context is usually expressed by a Greek or a Latin letter,

In a similar way, we do not want to use very complex visualizations. We visualize something very simple, almost generic. The specifics are introduced by other words used in the context. A good PAO visualization often encodes 5 words, which is approximately the number of keywords required for a quality google search. Once all 5 words contribute some of their meaning to the visualization, it cannot be very simple anymore. The complexity is an aggregate of chunking – not an intrinsic quality of the visualization.

Use chunking

The absolute beginners in memory arts and speedreading are taught to generate one visualization for one word, for all the words they need. Then they are taught to chunk the words in pairs so that the resulting visualization is less random. It takes a couple of years of practice to generate visualizations from four or five words very fast. And then there is a soft pause.

We can in theory use 7 words per visualization. From what I know, Jordano Bruno used this method. Astrological visualization with a planet per word. Only it is hard for most of us. Our visualizations become slow and complex. There is no return on investment. Two visualizations of 3-4 words are easier for most of us than on visualization of 7 words. Even after a decade of training, very few memory champions go beyond chunks of 4 or 5 words. At the same time, no memory champions I know of use chunks smaller than 4 words.

Simplicity can be an illusion

In physics, the most complex ideas are often expressed with very simple equations. It takes time to understand that each variable in the equation encodes a lot of meaning. Hawking’s formula for the black hole entropy is S=A*(kc^3)/(4Gℏ) where A is the area of the event horizon, c is the speed of light and k, G, and ℏ are the Boltzmann, gravitational and reduced Planck constants, respectively. If for a second you think that the formula is simple, try to understand the nature of each constant in it.