The Localization-Removal Test
Strip the crutch from an explanation to see if any real understanding remains
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 4
- Confidence
- 62%
Hoel's diagnostic strips an explanation of its most-used crutch and checks what survives. In neuroscience the crutch is location: experts explain a phenomenon by naming the brain region that lights up. Hoel would ask them to explain the same phenomenon again without referring to location. For early sensory processing (edge detection) a real mechanism remains. But once you move into the deep areas of the brain and remove localization, 'there's almost nothing to say.' The test generalizes: any explanation that collapses when its default shortcut is banned was never a mechanism, only a re-description. Swapping all the locational tags around would change nothing about how the system feels or works, which signals the explanation carries little real information.
Origin
Erik Hoel developed the habit in graduate school, where he describes being 'a menace' who repeatedly asked prominent neuroscientists to explain a phenomenon and then re-explain it without referencing where in the brain it happened.
Core principles
- 01Knowing where something happens is not knowing how it happens
- 02A real explanation survives the removal of its favorite crutch
- 03Localization can masquerade as mechanism
- 04Depth is revealed by what an expert can still say when the shortcut is banned
How to run it
- 1
Get the explanation
Ask the expert to explain the phenomenon as they normally would, in full.
Pro tip Let them use their standard vocabulary first so the crutch reveals itself.
- 2
Name the crutch
Identify the shortcut the explanation depends on. In neuroscience it is location; in other fields it may be a label, a correlation, or a black-box term.
Watch out The crutch is often the most confident-sounding part of the answer.
- 3
Ban the crutch and re-ask
Request the same explanation again, forbidding the crutch. 'Tell me how it happens, not just where.'
- 4
Measure what remains
Assess how much genuine mechanism is left once the shortcut is removed. Little or nothing means the original was re-description, not understanding.
Pro tip Try the swap test: if you could shuffle all the location tags and nothing meaningful would change, the explanation was hollow.
In the wild
Hoel describes asking neuroscientists to explain a high-level cognitive phenomenon, then re-explain it without saying where in the brain it occurred. For primary sensory processing like visual edge detection, a real mechanism survives. But for the deep areas of the brain, once localization is removed 'you'll notice there's almost nothing to say.' Much of what sounds like explanation is really a map: 'this area lights up,' 'damage here breaks that.' Hoel's swap test drives it home: imagine relabeling which region does auditory processing versus something else. You would feel no differently about the system, which is a strong indication the explanations were never meaningful.
→ Reveals that a large share of confident neuroscience explanation is locational re-description rather than causal understanding.
Common mistakes
Mistaking a map for a mechanism
Treating 'this region activates' as an explanation of how a phenomenon works. Location tells you correlation and lesion effects, not the causal process.
Applying it where localization is the answer
For well-understood early sensory processing, location does track a real mechanism. The test is meant to expose hollow deep-brain claims, not to dismiss genuinely mechanistic ones.
Is it for you?
Best for
Anyone evaluating scientific, technical, or expert claims and wanting to test explanatory depth.
Not ideal for
Early sensory or well-understood mechanistic systems where localization genuinely maps onto a known process.
From the transcript
“I would ask them to please explain it again but don't refer to location right don't just tell me where in the brain this is…”
“you could swap around all those locational tags... and you wouldn't feel any different about the system and that's a good indication that your explanations…”
From the episode
A Guide To The Fundamental Mystery Of The Mind - Erik Hoel - #663
Erik Hoel