1. Result: Blind adults show low agreement on the common color of objects (e.g. strawberry, stop sign = red). But blind & sighted make the same predictions for novel object color, based on knowing kind (natural/artifact) & its relation to color (e.g. color important to function) pic.twitter.com/qXDEDIk65n
— Judy Kim⁷ (@joodykeem) August 16, 2021
3. We argue: people develop causal-explanatory “intuitive theories” of color regardless of color experience. These theory-like understandings are, perhaps counterintuitively, *more* likely to be acquired than arbitrary associations (e.g., which color goes with which object).
— Judy Kim⁷ (@joodykeem) August 16, 2021
5. Exactly how color info is acquired is an open Q. Language is part of the answer, ofc. But importantly, to the extent that lang communicates about color, the “why” is acquired even when the “what” is not. So the key ? is the role of language in the development of these theories
— Judy Kim⁷ (@joodykeem) August 16, 2021
There's more to the tweet stream, but perhaps it would be more direct simply to download the whole paper. That's what I'm going to do.7. Another future direction is the role of pragmatics (thinking about what someone w/ different experience would say, e.g., a blind person thinking about sighted or vice versa).
— Judy Kim⁷ (@joodykeem) August 16, 2021
Also, a challenge for AI: whether “understanding” & not just that strawberries = red can be learned
No comments:
Post a Comment