You stand up with a clear purpose, walk to another room, and arrive with nothing. Not a vague memory of the intention, nothing. Then you go back, and it returns on the way.
This has a name in the literature, the location updating effect, and it is one of the few cognitive phenomena everyone has personally run the experiment on.
What the original studies did
Gabriel Radvansky's group at Notre Dame ran the core work through the 2000s and early 2010s. Participants moved objects around a virtual environment, carrying an item from one table to another. Sometimes the destination table was across the same room. Sometimes it was in the next room, through a door. The distance travelled was matched.
Memory for the carried object was worse after passing through a doorway. The same pattern showed up when the rooms were real rather than virtual, and, more strangely, when participants merely imagined walking through a doorway.
The explanation is the event horizon model. The mind does not store experience as a continuous stream but chops it into events, and it holds a working model of the event currently in progress. A doorway is a strong cue that one event has ended and another has begun. The old model gets filed, a new one is opened, and anything that lived only in the old one becomes harder to reach. It has not been erased. It has been shelved slightly early.
That the effect survives imagined doorways is the detail that matters most, because it rules out anything about physical movement and points squarely at how the mind segments experience.
The part usually left out
The story circulates as settled fact. It is not.
In 2021 a group at Bond University, publishing in BMC Psychology, ran four experiments across immersive virtual reality, video, and real physical rooms, and largely failed to find the effect. Their participants simply remembered the objects. It appeared only after the researchers added a concurrent working-memory load, and even then as a modest increase in errors rather than the dramatic blanking the popular version describes. Later work in virtual reality has similarly found that it depends on how the doorway is presented and how the participant moves.
The honest summary is that boundary-driven forgetting is real, that it is smaller and more conditional than the folklore suggests, and that it shows up mainly when you are already loaded. Which, if you think about the moment you actually walk into the kitchen and blank, is a reasonable description of ordinary life. You were not holding one clean intention. You were holding a half-finished sentence, a thing you were annoyed about, and the intention.
The mechanism you can rely on is not the doorway. It is working memory being small, and an unrehearsed intention being fragile.
Event boundaries are not only doorways
Segmentation happens at any strong change of context, which is where this connects to everything else about attention:
- Switching apps or tabs. A functional threshold. The cost of resuming is well documented as attention residue and context switching.
- Being interrupted mid-task. The interruption is the boundary, and the reason resumption is expensive out of proportion to the interruption's length.
- Finishing a meeting. The intention formed at minute forty is competing with the boundary at minute sixty. This is most of the argument for writing actions down before standing up, in how to take meeting notes.
- Changing rooms to work. The same effect that loses your intention also usefully clears the old context, which is why relocating to work can help. Boundaries are not purely a cost.
Note the flip side, because it is genuinely useful: the reason walking back into the original room recovers the thought is that the environment is a retrieval cue. That is context-dependent memory, and it is a real technique. Retracing your steps works.
What to actually do about it
The lesson is not to avoid doorways. It is that an intention held only in your head is held in the most volatile storage you own, and any boundary can knock it loose.
Externalise the intention before you cross the boundary. Say it out loud, or write it down. Two seconds. The whole argument for the capture habit is that thoughts are cheap to catch and expensive to reconstruct.
Make capture faster than the walk to the kitchen. If catching a thought takes unlocking a phone, finding an app and choosing a folder, the thought loses. This is why quick-capture friction is the single most consequential property of a notes system, more than search or structure or anything else. See how to capture a note.
Carry the object, not the intention. If you are going to the other room for a charger, take the device. Offloading the intention into a physical prompt is the crudest and most reliable form of external memory.
Reduce the load, not the doorways. Since the effect is strongest under load, the real fix is having fewer unresolved things resident in your head. That is what an inbox is for: see the note inbox.
Why it is worth knowing anyway
Even in its deflated form, the doorway effect is a useful thing to have a name for. It reframes a moment people quietly find alarming, especially as they age, as a normal property of how the mind files experience rather than evidence of decline.
Forgetting why you walked in is not a memory failure. It is a filing system closing a folder slightly before you were finished with it.
More in how memory works, cognitive load theory, and attention and focus.