What crows actually do around dead conspecifics
When American crows encounter a dead crow, they reliably do several things. They gather. They scold — producing mobbing-style alarm calls directed at the area around the corpse. They sometimes approach the body cautiously and inspect it. They occasionally bring objects to the location, though this is rare and not well-characterized. They subsequently avoid the area for some period of time. Kaeli Swift's PhD research at the University of Washington (in the Marzluff[1] lab) characterized this behavior in carefully controlled experiments where dead crow specimens were placed in areas with known crow populations. The behavior is real, reproducible, and consistent across populations.
Funerals, in the human sense, are commemorative rituals.
Why 'funeral' is the wrong word
Funerals, in the human sense, are commemorative rituals. They involve grief, social bonding around the loss, often religious or symbolic meaning, and a specific behavioral repertoire (eulogy, burial, mourning attire) that varies by culture. Crows do none of this. What crows do around dead conspecifics is more accurately described as a threat-response behavior. The gathering and scolding are consistent with a hypothesis that crows treat dead conspecifics as a danger signal — the location killed one of us, therefore the location is dangerous, therefore we should warn and remember. That's a different behavioral phenomenon from grief, even if the surface appearance has the dignity that human observers project onto it.
What Swift's experiments specifically showed
In Kaeli Swift's controlled experiments, dead crow specimens elicited stronger mobbing responses than dead non-corvid specimens (like pigeons), and stronger responses still when the dead crow was paired with a person holding it or with a hawk specimen present. The interpretation: dead crows function as a 'place is dangerous' cue, and the response intensifies when there's a plausible agent (human or predator) that could have caused the death. Crows also exhibited some learning — they were more wary of locations and individual humans associated with dead-crow encounters in subsequent observations. None of this contradicts the threat-response interpretation; all of it complicates the funeral framing.
Where the grief question actually lives
Threat response and grief are not mutually exclusive. A crow could be both treating a dead conspecific as a danger signal AND experiencing some form of emotional response to the death. The grief question requires evidence the threat-response framework doesn't address: physiological measurement of stress hormones in crows that recently lost a known associate, behavioral signs of prolonged change in surviving family members (reduced foraging, altered social patterns, altered vocal output), or comparative evidence across species. Some of this work has been done in other corvids (notably ravens), with suggestive but not conclusive results. For American crows specifically, the empirical grief case is thin. The threat-response case is strong.
Why anthropomorphism cuts both ways
The standard scientific worry about anthropomorphism is over-attribution: assuming human-like emotional states in animals based on superficial similarity. The reverse worry — anthropodenial, in the philosopher Daniel Dennett's term — is also real: assuming animals lack experiences they may actually have because we can't measure them. The honest scientific posture sits between: don't attribute grief without evidence, don't deny the possibility of grief because we can't yet measure it, and stay calibrated about what specific findings establish versus suggest. Popular coverage of crow funerals usually fails this test by reaching for the funeral framing without engaging the threat-response alternative.
What this means for AI bioacoustics
Dead-conspecific responses involve mobbing-style vocalizations that show up in the space as members of the mobbing cluster. The acoustic features of dead-crow scolds don't differ dramatically from mobbing alarm calls in other contexts. This is informative: it suggests crows aren't producing a special grief vocalization, but rather deploying their existing alarm repertoire in response to a specific kind of threat signal. The vocal repertoire isn't categorically expanded by the dead-conspecific context; it's contextually deployed. That observation comes from contemporary embedding-based pipelines and aligns with Swift's behavioral findings.
The honest summary
Crows gather around dead conspecifics. They produce alarm calls. They show evidence of learning about associated danger cues. Calling this 'funerals' or 'grief' overreaches what the controlled evidence establishes. The behavior is interesting enough on its own terms without the anthropomorphic frame. CrowLingo's editorial floor: when scientific careful work shows something specific, we report the specific. When popular framing reaches further, we name the gap. The 'crow funeral' story isn't false — it's just less precise than the empirical case warrants, and the precision is what makes the research interesting in the first place.