Amazing work @maorschreiber.bsky.social @francoisstock.bsky.social @nfaivre.bsky.social and @liadmudrik.bsky.social for leading this project, thank you!

thank you!

do you indeed think it's not penetrability?

This work was co-authored by the wonderful and social-networkless Nataly Davidson Litvak, and supervised by the brilliant @liadmudrik.bsky.social

Taken together, our findings suggest that semantic primes can modulate visual perception — a good case for cognitive penetrability.

We think this adds a fresh contribution to the debate, while avoiding some of the common pitfalls in past studies. Let us know what you think!

In a final experiment we showed the opposite – priming the illusory shape (“triangle”) facilitated detecting the illusion,

so a semantic prime could also enhance the illusion.

In three studies, priming Pac-Man both via images and via words reduced detection of a Kanizsa shape by roughly a half (comparable to the level of reporting a shape in our control, when no shape was even there).

We formed the illusion using Pac-Man shaped items, and hypothesized that priming the idea of Pac-Man can bias people towards seeing a bunch of separate Pac-Men chatting instead of the single illusory contour in between them.

So we took the Kanizsa illusion and tried to affect it using prior knowledge.

In the Kanizsa illusion, a shape emerges from imaginatively filling in gaps between separate items.

In this debate, visual illusions have a special role.

They present a “wrong” perception, but knowing it’s wrong doesn’t make the illusion go away – suggesting that perception may be sealed off from cognition.

Cognitive Penetrability sounds like a solid 90s Schwarzenegger film but is in fact a longstanding debate in cognitive sciences –

do our thoughts, beliefs and feelings affect how we interpret what we see, or do they actually make us see the world differently?