Method of loci vs linking

I have been using linking without knowing it was a commonly-used mnemonic technique, and I have had some success with it. I would repeat each link in the list independently, and recently have noticed that recall is better if on top of that, I also drill the entire list, after memorizing each independent link. This linked list has always been of phrases, not images.

I recently decided to give the method of loci a go, and I’m pleasantly surprised with the results, not so much with how easy the sequences are (which is pretty much what I expected), but also with how ‘sticky’ the information I put in a loci is. I tried the method of loci with numbers (even if the usual recomendation is to use pegs), and with a list of 6 items that I had some trouble remembering. I managed to remember all 7 numbers, and 5 out of 6 elements in the list; I doubt I’ll forget the 6th element again with sufficient maintenance of that loci, though I might adopt a peg system for numbers instead.

I have tried to find some comparisons of linking vs the method of loci: it’s obvious that memory palaces are more versatile, though harder to use. I don’t think the amount of loci available will be a problem, and while I have been linking phrases, I know it would be easier to memorize if I convert the phrase to something visual.

What I’m concerned about, is to employ the method of loci for things that are better suited for linking. Linking has the advantage of being easier to learn, but with some practice this difference disappears. Also, the method of loci has the obvious extra overhead of having to use/create a memory palace, and create a scene (though scene creation could be equated with the overload of turning linked phrases into visuals), but this extra effort could be mitigated by the fact that linked lists are harder to remember than memory palaces. So if there’s people who use both linking and the method of loci, when do you use one or another?

Often they’re both used simultaneously, you have a locus (location) in a memory palace and in it you have linked a person with an action and an object, or you have three objects linked together. You can also link multiple loci together, making the associations stronger and easier to remember.


Edit: Also see this old post. Interesting stuff in there.

1 Like

I understand they can be used together; however, my interest is not to compete in memory competitions, so given the power of mnemonics, it seems that the way I have set up the method of loci will fulfill my needs (I picked a fictional planet based on earth for permanent knowledge, and an appartment I used to live in for temporary knowledge, and using 10 places inside buildings). For this same reason is that I don’t think I’ll ever create a peg system for numbers with more than 100 pegs, maybe even 10 pegs will be enough, as the longest numbers I’ll have to memorize are 10 digit phone numbers.

In the post you linked, these are the listed differences:

  • The method of loci takes more time to create... then someone else said he eventually developed his skill into the opposite
  • Linking is less resilent to forgetting one item in the chain
  • Linking always requires to go through the entire list; though this can be fixed by adding pegs every X item, but this adds complexity
  • Linking is only reliable when you do not have duplicates, if you use words
  • Linking do not require loci

I just created a simple 3D object for a 7 item list, without an attached location, and it seems to be working fine… it seems simple visualization works just as well for small lists, not even linking being required.

I strongly favour the method of loci for large amounts of data because I can jump to anywhere, go forward and backward, and it is resilient. The main reason is that I can then add layer after layer of further detail without losing track. However, I then use linking within the method. I have a number of long journeys set up, both in the house and streets I can walk around, and on hand-held devices, based on indigenous memory aids. They are the same principal - a sequence of physical places for loci.

As an example, I have the 405 birds for my state (Victoria, Australia) memorised in taxonomic order. I have the 82 families linked to a sequence of locations. I then have a story - effectively a linked list - for the birds within that family, unless there are more than 4. In that case, I add another set of loci, each having up to four birds in it. So I am using small linked lists within the journeys. Those small lists are essentially stories, much like indigenous cultures use. Then I add more and more complexity to each story as I add in ID and habitat and any further details about each bird.

I am doing similar things for all the countries in the world, and timelines for all of prehistory and history. In the last two cases, I want to add in historical events constantly, and they will come in between previously memorised events. I walk through time as I trot around the block with the dog. Linking wouldn’t work there at all.

Hope I have made sense. Like you, it is not competitions which interest me but memorising masses of useful information. But none of it is temporary. I want to remember it all for ever!

Sounds like an interesting experiment. I look forward to hearing more. All the very best with your memorising!


I understood perfectly.

However I’m about to drop the method of loci, and stay with linking after transforming the ideas into visuals, and pegs.

Pegs have been proven useful when memorizing numbers, and while the method of loci has worked with it, I only have memorized 2 number sequences.

The problem I have had with memory palaces is that it’s time consuming. Information that is usually not sticky turns out to be very sticky, and linking works great for lists; but then I have problems with remembering what was in each room, and in which part. In some cases I completely forget an item, in some cases I’m not sure in which room the item is. I started with the cube method (the one that provides 10 loci), but later read that the original creator of the technique additionally had an item associated with each location, besides the item he wants to remember; that implies using a lot of time per room, which makes sense if you are going to reuse the room for competitions, or if you have a memory palace for temporary storage, so that technique is useless for simply memorizing large amounts of info, as there are more efficient ways of doing it than using the loci method and thinking about 10 unique items plus the items you want to remember per room. Basically, it is a technique to gain a burst of speed with the loci technique, at the cost of much greater use of time pre-competition.

I have tried the method of loci in it’s vanilla variant, and it has worked great to a degree; with pegs it has not failed me yet. With normal objects, I have forgotten about 60% of the items, but note that I only use memory palaces with items that I could not remember any other way, so it’s still a win.

However, I suspect that simple visual linking will provide better results, and that the method of loci as a whole serves better as temporary storage. Of course, the problems of linking are known when you don’t use the method of loci, but for the amount of data people manage everyday, the loss of a link can be easily fixed by chunking, and the interference is only likely to happen for very long lists, or if you use similar visual images for the same lists, but it’s still less likely to be interference than memorizing text.

You also have to take care and make sure rooms with the method of loci are easy to diference; since a house’s rooms, for example, are likely to be very different, there’s little risk for confusion, but what if you pick a school as a memory palace? all classrooms are bound to look similar. All these small details and considerations that one has to take with the method of loci, forces a larger amount of time spent in memorizing something that could memorize almost as well with visual linking, unless you have pre-made your rooms, but that limits the efficiency of the technique.