• Search |
• Edit this page

The relationship between language production and understanding

Could it be that language production depends on language understanding? Let's make a few assumptions:

• We produce language by the seat of our pants. We don't know exactly what word we'll utter until immediately before it needs to come out.
• The cognitive origin of a sentence is a pattern of activation in some sort of semantic network (call it original pattern).

Given these, perhaps we create language like so:

1. The combination of the original pattern and a desire to vocalize it activates a candidate sentence structure.
2. That structure lights up a part of speech that we need.
3. The pattern for that part of speech, in combination with the original pattern lights up a candidate word.
4. We utter the word.
5. We use our language understanding mechanism to interpret what that little bit of sentence means. This gives us another pattern in the semantic network (call it conveyed pattern).
6. We match the conveyed pattern against the original pattern, obtaining a boolean difference (call it remaining pattern.)
7. Assuming the difference is larger than we'd like, go back to the beginning, and use the remaining pattern instead of the original pattern.

I guess in step 6, we would continue to use the real original pattern, not the new remaining pattern.

The obvious way to find out would be to see if there are people who can produce language but not understand it. Still, I wonder if there is truth to this? It may be that we can't really process language fast enough to keep up with the production. And I'm not necessarily suggesting that we use our ears to process the language. It may be it doesn't leave the brain--it just uses the production machinery.

This might be testable with neuro experiments. It might also be worth doing an experiment that played loud, bassy music or something while someone spoke to see if it affected their ability to produce. Ideally we'd want to dampen their mental language processing too. Perhaps with interfering speech. It'd be hard to prevent them from processing what they are saying without introducing other interference. And I can't think of a way of controlling for that.