John Marzluff’s team at University of Washington wore specific masks while trapping and banding crows on campus. then walked around wearing either the “dangerous” mask or neutral control masks.

crows that were trapped mobbed the dangerous mask immediately. expected.

crows that were never trapped also learned to mob the dangerous mask. they learned from watching other crows’ reactions — no direct negative experience needed. the knowledge spread across the population. young crows born after the trapping learned to mob the mask from adults.

recognition lasted 5+ years. crows were still mobbing those masks years after the original events. and they stayed specific — didn’t generalize to “all humans dangerous,” just that particular face.

neural studies showed crows use dedicated social cognition pathways similar to primate facial recognition. specific brain regions activate for threatening vs. neutral faces.

sam’s beach crew is the same phenomenon, inverted. marzluff: negative association, 5+ year memory, social teaching, specific recognition. sam: positive association, 5-month memory (so far), social teaching, specific recognition of face + vehicle.

one study terrorized crows to prove they remember. the other feeds them and documents the same mechanisms.

makes you wonder: can crows update their assessments? if the “dangerous mask” person fed crows for months, would they revise? if sam disappeared for years, would the memory fade or persist like the threat memories?

and how do they integrate information? face + vehicle + location + behavior = contextual model. they’re not just recognizing sam, they’re predicting his behavior based on historical patterns. young birds learn this from adults somehow — vocalization? modeling? active teaching?

marzluff’s research gave us hard evidence of corvid facial recognition and cultural transmission. but the cost was borne by the crows — trapping, stress, years of being mobbed for wearing a mask. sam’s accidentally generating similar data points by being consistent with peanuts.

worth asking whether there are better ways to study wild cognition that don’t require terrorizing the subjects first.