- thought I had based on ↑ ~ avoid using the same stimulus for different intended reactions
- as in:
- when you learn apple=tofe7a, your brain has to sometimes react to "apple" with "apple", and sometimes with "tofe7a"
- (note: this really just reframes believability)
- when you stick to don't translate, you kind of overcome this by
- sort of just auto-completing thoughts in your target language, fitting in tofe7a because it has the right gestalt
- sort of storing a different gestalt for tofe7a that's likely close to apple but not quite equivalent to it, since no word actually translates to anything directly
- when you learn apple=tofe7a, your brain has to sometimes react to "apple" with "apple", and sometimes with "tofe7a"