Feynman is not talking about subvocalization in that video. He's talking about seeing and hearing. According to psychology, humans have two sets of "short-term

Feynman is not talking about subvocalization in that video. He's talking about seeing and hearing. According to psychology, humans have two sets of "short-term memorybuffers"; spatio-visual, and auditory. It is perfectly possible to do two tasks at once, as long as you use different "buffers". Like juggling and singing. But to be able to comprehend text, it needs to go through the linguistic centre. It doesn't matter if it goes through the spatio-visual or auditory sensors. Subvocalization is created by the linguistic centre. Simple words or numbers do not need to go through the linguistic centre. So it is perfectly possible to comprehend text and at the same time count if you "see"/"hear" text and "hear"/"see" counting. But you cannot comprehend two advanced texts at once.