Reusing a Palace Without the Ghosts

One route, tested, spaced out, holding steady — that's where the last guide left off. So for the next list, you reach for the same palace: familiar, every stop already known. Why build a new one from scratch?

Walk it, and something's off. The kitchen counter still has last week's coffee cup melting through it, the doorknob jammed with keys from a list you don't need. You wanted an empty palace — instead it's full of stuff that should be gone.

That's not a sign you did something wrong — it's one face of a well-studied memory effect, and it has a name. Two, actually.

The short version: old images linger in a reused palace because of interference — old and new material competing for the same retrieval path, worse the more alike they are. Memory athletes call the leftovers ghosts, and it happens to champions too. The fix isn't scrubbing harder — it's letting images fade on schedule, keeping different kinds of information in different palaces, and spreading new lists across stops instead of stacking them. The rest of this guide is those three ideas, one at a time.

Old and new memories compete for the same space

Psychologists were studying this long before anyone called it a memory palace problem. John McGeoch, in 1932, studied retroactive interference — new learning getting in the way of recalling something older. Benton Underwood, in 1957, studied the reverse: proactive interference, where earlier learning makes the newest thing harder to pull back. McGeoch's argument was blunt: forgetting isn't mostly decay, it's other things crowding it out.

That's roughly where the field still sits, though not closed — some researchers think plain decay deserves a second look too. What's held up more consistently: similarity makes interference worse — the more two things resemble each other, in content or context, the harder they interfere. None of it was tested on memory palaces — just word lists in a lab. Treat it as the general shape memory takes, not a study of your palace.

The community has a word for it: ghosts

Palace practitioners noticed this long before anyone cited a paper, and gave it a name — ghost images: the leftover picture from a previous list, still hanging around a stop you meant to reuse. And it's not a beginner problem. Alex Mullen, three-time World Memory Champion, has talked about old-deck images lingering when he returns to a palace. Nelson Dellis, four-time USA Memory Champion, describes a real risk of "confusing things" — echoes of the old set intruding. If ghosts find the champions, they're not a verdict on your technique.

Let old images fade instead of forcing them out

The instinct is to scrub a palace clean before reusing it — walk through and will the old images away. Mullen runs the opposite: he's said he doesn't actively clean palaces, just leaves one alone and lets images fade — by his own estimate, a matter of days — before trusting it with something new.

Set that next to the last guide's advice — on the surface they pull in opposite directions. It said: review a route you want to keep, spaced out, because reviewing is what keeps an image from fading. This is the mirror image — a palace you want to reuse gets the opposite treatment: stop reviewing it. Let it sit, and the fading that review normally prevents is exactly what clears the space. Review keeps. Rest clears. Same mechanism, different goals.

There's no universal number of days — start with something in the range of a few days to a week, then let the ghost itself tell you: still showing through when you return means it needs longer.

If a ghost keeps bleeding through even after a rest, rebuild the new image so it looks nothing like the old one. It's a different problem than the sixth guide's blank stop, but the same fix: remake the image rather than force it to hold.

Give each palace one job

A second move: stop asking one palace to do two jobs. Some memory coaches recommend splitting palaces into two kinds — temporary ones you expect to overwrite, and permanent ones reserved for what you actually want to keep. The point isn't that one kind is smarter; it's that you stop confusing yourself later — a permanent palace never gets casually overwritten, and a temporary one never needs babying.

Spread new lists across stops, not onto the same ones

If reuse is unavoidable, there's a third lever: where inside the palace things land. In 2014, psychologists William Bass and Kimberly Oswald had people memorize several same-category word lists — fruit, in this case. With no technique, recall on the last list fell sharply. With the Method of Loci — each list keyed to its own set of stops — the drop was much smaller. The study never tested recycling stops across lists, but it fits the pattern above: same spot, similar content, close to worst-case for interference. Separate real estate is what held up.

It's also why competitive memorizers rarely lean on one palace. Many memory athletes report rotating through a pool rather than reusing one back to back — Mullen has mentioned working through around ten. By the time you're back to the first, whatever was there has usually faded.

Check yourself before you move on

• Before reusing a palace, have you let it sit untouched for a few days, not just assumed it's clear?
• Are you sorting palaces into "temporary, expect to overwrite" and "permanent, keep separate" — or treating every one the same?
• When two lists are similar, are you giving them separate stops or separate palaces, instead of stacking them together?
• If a ghost keeps bleeding through, have you rebuilt the new image to look nothing like it — not just tried to remember past it?
• Do you have more than one palace to rotate through, so none gets reused twice in a row?

Mostly yes, and reuse won't cost you the accuracy you worked for. A no or two just points at which habit to add before the next list.

Where to go next

One palace, reused carefully, gets you further than you'd think. But memorize more than a handful of lists and you'll end up wanting several — which raises questions this guide didn't touch: how many is enough, how to build them so they don't blur together, and when a new palace beats squeezing one more list into an old one. That's next.