Is it like a rough inference of what’s being said based on mouth movements, or is it more precise somehow? Would it be a mistake to think you knew exactly what was said by reading lips (even if you were good at it)?
Is it like a rough inference of what’s being said based on mouth movements, or is it more precise somehow? Would it be a mistake to think you knew exactly what was said by reading lips (even if you were good at it)?
It’s inference based on mouth movements, but it isn’t as rough as it seems like - context plays a huge role on disambiguation, just like it would for you with homonyms that you hear. It’s just that the number of words that look similar when you mouth read is larger than the number of words that sound the same, since some sounds are distinguished by articulations that you can’t immediately see (such as [f] vs. [v] - you won’t see the vocal folds vibrating for the later, so “fine” and “vine” look almost* the same.)
Also, the McGurk effect hints that everyone uses a bit of lip reading, on an unconscious level; it’s just that for most of us [users of spoken languages] it’s just to disambiguate/reinforce the acoustic signal.
*still not identical - for [v] the teeth will touch the bottom lip a bit longer.