Help to Formally Update Mnemonic Classification and Nomenclature

Hi there,

I am a Neurology doctor based in London with a Masters in Medical Education. My research looks at mnemonic strategies and how they can be best incorporated into the undergraduate curriculum.

I would like to ask for your help to optimise how mnemonic techniques are classified and named. Hopefully, through this, we can develop a robust, updated mnemonic nomenclature and classification system. I believe this is very important to help further formal peer-reviewed mnemonic research. The last comprehensive classification system was proposed 40 years ago (Bellezza 1981).

Bellezza’s classification, compared to other mnemonic classification systems (Thompson, 1987; Oxford, 1990; Baddeley, 1999) is better because it separates mnemonics hierarchically into encoding and organisational mnemonics. Encoding mnemonics is the process by which info to learn is encoded to be more memorable, and organisational mnemonics refer to how that encoded information is structure. Normally encoded and organisation mnemonics are used concurrently but can be used in isolation if needed.

I’ve modified Bellezza’s to the flow chart below (added kinaesthetic mnemonics, moved link and story mnemonics from multiple-use to single-use).

If you could help answer the specific questions below this will really help (don’t have to answer all, some are fine :smiley:)

  1. Do you agree with the classification system proposed? - If you do (or importantly don’t agree), please elaborate. Any recommendations would be great
  2. Do you agree with the naming (nomenclature) given to mnemonic subtypes? - If you do (or importantly don’t agree), please elaborate. Any recommendations would be great
  3. What would you name a mnemonic where multiple visually encoded information is associated with a central image? (for example, see here - ignore the audio, just focusing on the picture). Initially, at face value, this is a chain-type link method, but the information is not chained sequentially where 1 linked to 2, linked to 3, linked to 4…, instead, it would be 1 linked to 2, 1 linked to 3, 1 linked to 4 etc, where all the information is linked to a single main image.

Any way you can help would be greatly appreciated!

Baddeley, A.D. (1999). Essentials of Human Memory. East Sussex: Psychology Press Ltd
Bellezza, F. S. (1981). Mnemonic Devices: Classification, Characteristics, and Criteria. Review of Educational Research, 51(2), 247–275. doi:10.3102/00346543051002247
Oxford. R. (1990). Language Learning Strategies: What Every Teacher Should Know. New York: Newburry House publishers
Thompson. I. (1987). Memory in language learning. In A. Wenden 8J. Rubin (Eds). Learner Strategies in Language Learning. (pp. 15-30). New Jersey: Prentic-Hall


One note, acronym is under abbreviation. Abbreviations can also be treated as simplification. Eg Demo and demonstration, Dr. and doctor, ST and street etc

@thinkaboutthebible has done work on the classification and nomenclature of memory systems:

Myself, I believe that many systems can be combined: Method of loci be combined with song/music, a grid like ordering system, pegwords etc.

  1. What would you name a mnemonic where multiple visually encoded information is associated with a central image? (for example, see here - ignore the audio, just focusing on the picture). Initially, at face value, this is a chain-type link method, but the information is not chained sequentially where 1 linked to 2, linked to 3, linked to 4…, instead, it would be 1 linked to 2, 1 linked to 3, 1 linked to 4 etc, where all the information is linked to a single main image.

This is really a version of the method-of-loci - the locus (background) here is the primary image. To me, there isn’t much distinction between a background and image - a background can represent information as well.

Another thing not touched upon in the chart is how individual images can be constructed - two separate images can be linked together, but also you could just use PAO (Person-Action-Object) or SEA-IT (Subject-Enhancement-Action-Item-Terrain) to create an image that represents up to 5 things.


Also not mentioned here, are physical memory objects. Many aboriginal memory techniques (see “Memory Craft” by Lynne Kelly) use actual physical objects, e.g. a stick with symbols cut in it, to memorize vast amounts of information. The symbols are not writing, they are like memory pictures with which information is associated.

Your own body, or fingers of one hand is such an example.


Very true, thank you

1 Like

Agreed, but good to have a formal classification system as most people who aren’t well versed in mnemonics are likely to just use one

1 Like

Wouldn’t this just be keyword mnemonics combined with links

1 Like

I see your point, but I thought the location method is an adjunct device where you can encode information onto it, so if you remove the information to learn, you can reuse the location. I am thinking about what to name a technique without a location, which can only be used once for that particular information. Not always is background information encoded, i.e. image central keyword mnemonic encoded visual, with 5 visual pieces of info attached to it, you would just need to recall the image itself. I do get your point how this could be a location but not fully convinced

1 Like

Yup aware of Lynne Kelly, great stuff. I wanted to focus on internal mnemonics, not external for purposes of cognition. I guess in this sense would be a external method of loci? Using hand/finger etc I guess is just an elaboration of loci?



Hi there, @Niten mentioned your post on analysing visual mnemonic systems - amazing work! Looking through the published literature I have not come across a classification system similar to yours. I was hoping if you would be able to look at the proposed classification above through your lens of work on visual memory systems to see if this can be updated/changed in any way. Thank you



On one hand, I like the idea of categorization. On the other hand,

  • I disagree with some of these “splits” for example:
    • single use versus multi-use is about the intent the user has, and isn’t necessarily intrinsic to the technique itself.
  • These can all be mixed together in a way the blurs the lines-- and often that’s more effective.
    • For example: To recall a book I have: created an acrostic out of the chapter titles in such a way that the acrostic describes a story. The story includes a pegword character and I visualized it within a location and includes various encoding types including action and word encodings.

I personally think of it like an additive process of memorable elements. You keep stacking on, so you get more of a list of elements. When I quiz myself on something I’ve memorized… say a list of 200 things. I often forget a few things. When that happens, I add more mnemonic detail to make it stick better. It’s a personal process because you’re trying to link to things you already know, and can already recall. Which is of course different for everyone. @metivier Talks about this with “KAVE COGS” and so do many other other practical memory authors.

For example, when Ron White did that “Memory Wall” project (recalling all the US soldiers to died in Afghanistan) he had issues between similar spellings and added pegs for different similar spellings. (But this was already within a method of loci system)

Recently I’ve been playing music, and as a result, I’ve been adding sound to my memory palaces. For example, to recall “side by side comparison” I have an image of the cover of the film The Usual Suspects. Now, I’ve added the sound so of the suspects nervously shuffling under the spot lights and that emotional sensation of tension and suspense.


@moo Thanks for your input. Just to clarify, the ‘usage type’ is in regards to whether the same cognitive construct can be reused or not. e.g. to learn a deck of cards, you use a location method, which doesn’t change to learn deck order. After your done, if you wanted, you can use the same location method to learn a new deck order - so there is the potential to reuse the method. Single-use method would be if you learned so that 1st card linked to 2nd card etc, or created an acronym, or song etc,. It can only be used once to learn that particular deck order. To learn another deck order, you’ll have to create a new mnemonic. Of course you can continue to learn from the single use mnemonic :slight_smile: How would you consider amending the terms to make it more apparent?

Totally agree, there can be lots of overlap between mnemonic types, the categorisation in itself does not mean mnemonics can’t be mixed, rather how they are related. The organisational system is usually how the mnemonics are stored, which from your example fits well as you are using pegword and location mnemonics as a storage container, and to learn the book info using single-use acrostics/stories.

Let me know if agree/disagree :slight_smile:

1 Like

@swiftdeck, thank you for your compliment. I caught your post this morning and as a result spent the rest of the morning fine tuning the glossary on my github site, which has been my passion for over a year. Youll find a memory systems overview as well as a decision tree there that explored categorization and usage of the terms. I version the documents so, check there to make sure you have the most recent ones.

I’ll post another reply for some comments on your classification system and what not.

It’s extremely important to me to have a consistent systematic way of talking about the mnemonic field before I can talk about it and never found a good explanation so I made one. I spent many hours reading through posts here to get a sense of usage and examples as well as testing the systems for me. I just finished a historical overview to make sure I didn’t miss anything there and am satisfied that what I’ve done will help advance research like yours in the future.


Awesome work, nice to see a concerted effort to help formalise mnemonology! I’m not sure if you’ve read through Worthen and Hunt’s book, they talk about the varied terminology. Look forward to your thoughts on classification :slight_smile:

1 Like

Just came up with a nice term to encapsulate this; ‘focal peg’. I think this nicely draws in the fact that multiple bits of info is focused i.e attached onto the main peg. Was thinking about ‘anchor peg’, however this might be interpreted as the initial image being an anchor, and other info is sequentially attached to it in a chained manner which would not be right.

Also was thinking that multiple-use and single-use might be interpreted, incorrectly - thoughts on saying reusable or non-reusable instead?


I have some work to do today, but am excited to work with you on this, @swiftdeck.

Here’s some initial observations:

  1. Encoding is a popular word to use for what is really a translation of one language to another, whether numerical, visual, aural or other. Encoding assumes hiding of information as a goal but the goal is usually to bring the information into the visual realm so I choose to say translation. (I was a military cryptanalyst for a time and know about encoding systems.)

  2. Encoding mmemonics would be seen as my visual sentence creation where the data types are assembled and associated with appropriate translations. Sentences have various structures and organizations.

  3. Organizational mnemonics would be seen as my traversal systems being either narrative or rule-based (pegs) as you associate the various visual sentences.

  4. Link is a generic word for association and has many applications which should not be used as a classification but as a foundational skill.

  5. As you can tell, I somewhat agree but have many suggestions on Bellezza’s system. Reading through my documents will give you my recommendations. I’ll read that paper soon. It’s quite good I thought. The YouTube medical mnemonic video illustrates a visual sentence. A subject-action-item-terrain rich visual image. I recently explored the creation of that with the Snow White and the Seven Dwarfs topic here on the forum.

I did an analysis on number pegs, found in my repository, to determine what metrics to use when measuring the efficiency of the systems in my repository. It surprised me that I could come up with 92 different systems. I analyse them and rank them. That’s the most academic study that I’ve done and tried to be as precise in nomenclature, methodology, and result reporting as I could without any previous examples.

I’ll try to respond with more soon.

1 Like

Thank you for raising that, Niten. I have a PM from swiftdeck, and have raised this issue. I understand where he or she is starting from, but I am finding that I use physical components in all I do in my own experiments and with students - school and university levels. Working with educators (and, when possible) Indigenous colleagues, we are always using integrated systems because we draw on Indigenous techniques. I am getting a headache just thinking about how I might try to fit what we do within a classification system!

Challenges ahead for the academics working in this field!

1 Like

Reading through the comments, the over-arching theme seems to be that the previously proposed classification system is too restrictive. Taking this on board, I’ve aimed to improve the language further regarding classification, which I hope can now help quickly identify what the ‘structure’ of the mnemonic is based on its classification without getting too bogged down with differentiating subtypes, so that the classification remains flexible and adaptable to change. See updated classification below

Improvements: Improved language, instead of organisation, scaffold I think helps convey the structural element that the mnemonic provides. Changing multi-use/single-use to pre-made and ad-hoc covey more about how the mnemonic is generated. Based on recent recurrence about external memory aids (thanks @LynneKelly), felt best to create this in parallel with a pre-made scaffold (correct me if wrong).

Another nice thing about this structure is that its organisation helps you to understand what the mnemonic is about, i.e. the loci method is a mental, pre-made, scaffold mnemonic or, the story method is a sequential, ad-hoc scaffold mnemonic.

All thoughts welcome, let me know what you would change :slight_smile:


@swiftdeck, I applied my lens to memory systems for you. After 25 years of coding and analysis you realize that models don’t mean much unless you can find a way to test them. And software is a harsh test. Most software before Object Oriented coding in the 2000s got to a point where the size of the application made any updates impossible. The classification of the entities were mostly taxonomies of data types and led to difficulties as you tried apply several different entities to a real world problem. The phrase “death by taxonomy” pointed to a necessary change in the perspective of software models that concentrated not on traits but on actions or behaviors so that the model better reflected reality.

I ended up creating a rough OO model of mnemonic software as a cognitive process for the computer to understand the systems and data structures that would generate the categorization that I had grasped in my Memory systems overview. It’s a good fit. My glossary is a good fit. It is all consistent which was my goal.

It also helped my clarify a curriculum for education as the process would be built up from the building blocks that are modeled in the class diagram. And I’m a retired professor with a 183-page document to explain what it means so I can teach others. But it’s changed much over the last few years.

I use some graphics like the key-value Tinker Toy models you have as well. But I think to grasp the scope of what systems people use today, the taxonomy approach will never be sufficient unless you want just a general description to focus on traits for your research. Take a look at the topic I started for the software model:

That and the Memory systems overview should give you a good idea of the differences in our models. If you want, I also wanted to break down your medical YouTube illustration into my datatypes like I’ve done for many people here if it will help you see it as a visual sentence.

Having specific questions like you offered at the beginning post was a great way to have a dialog. Let me know what I can explain or help you see the differences as there are so many working pieces here.

Thanks for all the psychology references. I’m working my way through Worthen and Hunt’s Mnemonology which is wonderful, but didn’t see much to help me in Bellezza’s model. I’ll go through it one more time though.


Thanks @thinkaboutthebible for the comprehensive review so far :slight_smile: Looking at your diagram, things appears more similar than dissimilar, which adds to the importance as you said to be generalisable to an extent. There will always be subtle differences between people who independently create classification systems, but the fact they have good alignment is telling.

Glad you like Mnemonology, its a lovely read. In my head my YouTube Neurology animation videos are mainly analogy based, so in my mind did not classify them as a mnemonic method. This is based on Levin & Levin 1990 work looking at the directness of correspondence between new and known information - mnemonics are furthest apart, whereas analogies share similarities between new and known info. It would be interesting to see how a visual sentence can be made for the videos :slight_smile:

Scientific Mnemonomies: Methods for Maximizing More Than Memory on JSTOR