Do Ravens Recognize Humans? Four Years of Evidence

Yes. One study followed the same birds for four years and some of them were still reacting to a face they had seen go wrong once, in 2011, with nothing to remind them in between.

There are two separate pieces of research here and they answer two different questions. One is about a threatening face. The other is about a person who cheats you in a trade. Only the second one used real, uncovered human beings.

The short answer

What was tested

Whether ravens treat one human differently from another after a single bad experience with one of them

How fast they learned

Often on the first trial

How long it lasted

Four years in the mask study. At least one month, and in one bird two months, in the trading study

What it does not show

That a wild raven can pick your particular face out of a street. Both studies used captive birds, and one of them used rubber masks rather than faces

Four years from one bad afternoon

The long study was run at the Haidlhof Research Station in Austria and published in 2020 by Christian Blum, W. Tecumseh Fitch and Thomas Bugnyar. Sixteen captive ravens, kept in two aviaries, were shown people in matching clothing wearing one of two masks.

Control phase, October 2011

Both masks were presented with nothing attached to either of them, to see how the birds reacted before any training

Training phase, October to November 2011

One mask now appeared carried by a person holding a dead raven the captive birds had never met. The other mask appeared empty handed. Four pairings, alternating with unreinforced presentations

Test phase, November 2011 to October 2015

Both masks presented again, 28 trials, with no dead raven ever shown again

What was measured

Scolding duration, meaning the proportion of each presentation the bird spent producing alarm calls

The authors write that the ravens discriminated between the masks "often based on the first trial." That is the part worth sitting with. Not a training curve, not forty sessions. One look at a bad thing happening.

And then the tail on it. Some individuals were still separating the dangerous mask from the neutral one for "the next 4 years," with nothing during those four years to keep the association alive.

Why a mask instead of a face

The same reason the crow work used them. A mask can be worn by a different person each time, which means the height, the walk and the voice underneath stop being possible explanations. What is left is the thing on the front of the head. We went through this in more detail in the article on crows and human faces.

It also sets the limit of what the study proves. A raven learning to shout at a specific rubber face is strong evidence that ravens store and retrieve face shaped visual information for a long time. It is not the same claim as a raven knowing you personally.

Not every raven reacted the same way

All sixteen birds saw the same thing. They did not behave as one animal, and the paper is unusually direct about it.

Rank mattered

Dominance was scored with Elo ratings, and higher ranked birds produced substantially longer alarm calls. The paper reports this at p less than 0.001

Sex did not

Dominance, not sex, explained the differences between individuals

Upbringing mattered

Parent raised birds scolded more intensely than hand raised birds

Group size mattered

Larger groups produced less scolding per individual, which the authors describe as a dilution effect

Group changes showed up

When the composition of a group changed, individual scolding patterns changed with it

The practical reading of that list is simple. The bird making the noise is not necessarily the only bird that remembers. Scolding is a social act as much as an alarm, and a quiet raven at the back is not evidence of a raven that forgot.

The study that used actual people

The second piece of work, published in 2017 by J.J.A. Mueller and colleagues, dropped the masks entirely. Nine captive bred ravens were taught an exchange: hand over a piece of bread, receive a piece of cheese, which ravens rate as the better item.

The fair partner

Took the bread and handed over the cheese as agreed

The unfair partner

Took the bread and then ate the cheese in front of the bird

The test

Later, the same birds were offered a choice of who to trade with

The result

Birds that had been cheated refused the cheat and took their bread to a fair or neutral person instead

Time reports the duration as avoiding those people "for at least a month," and the summary of the paper is that ravens can remember, "after a single interaction sequence, who is worth cooperating with in the future." One bird held out for two months before it would deal with a previous cheat again.

Watching it happen was not enough

Two extra ravens were present but never traded themselves. They observed the whole thing.

Those observer birds showed no preference at all between the fair and the unfair person. Direct personal experience was what built the memory.

That is worth flagging because it cuts against the story people usually tell about corvids. In the crow work the alarm response spread through birds that had never been handled. In this raven experiment it did not spread to birds that had only watched. Two different species, two different tasks, two different answers, and we are not going to blur them into one folk claim about corvids warning each other about you.

Is the bird in your yard actually a raven

For most readers in the United States, probably not.

The Cornell Lab describes Common Ravens as occupying nearly any habitat in the Northern Hemisphere, but notes they avoid the eastern forests and the open Great Plains. More to the point for a garden: in larger towns ravens "are often replaced by American Crows, although they do occur in some cities including Los Angeles."

So if you are watching a big black bird from a kitchen window in an eastern suburb, the likelier answer is a crow, and that bird has its own article: do crows recognize human faces. We have also written up mockingbirds and pigeons, which between them cover most of what actually turns up at a basin.

If ravens really are visiting

This section is our judgment applied to what the sources say, and we are marking it as such.

1. The water is not why they came. Cornell lists the diet as carrion, small animals ranging up to an adult Rock Pigeon, arthropods, grain, and many types of human food including unattended picnic items and garbage. A basin is a drink on the way to something else.
2. Expect to find things in it. Ravens cache surplus food for later. A bird that stores food around your garden will sooner or later store some of it somewhere you did not want. Cleaning method does not change, only how often you do it, and that is in the guide to cleaning a concrete bird bath.
3. Do not run the experiment on purpose. The finding above is that a single unpleasant encounter can stay on the books for years. Chasing a raven off is a decision with a long tail, and there is no evidence you can take it back.
4. Watch the small birds instead of the raven. Cornell puts adult pigeons inside the size range ravens will take. If the songbirds abandon a basin while ravens are around, that is not a water quality problem and no amount of cleaning will fix it.

What we could not check

What we read in full

The 2020 mask paper in Frontiers in Psychology, the Time and Phys.org reports of the 2017 trading study, and the Cornell Lab species account

What we did not read

The 2017 paper itself. Every number we give for that study comes from the two reports, and we have said so each time rather than dressing it up as the paper

The gap neither study fills

No work we found tested wild ravens on the uncovered faces of specific people. The long study used masks. The people study used captive bred birds in a trading task. Neither is a raven in a garden knowing which neighbor you are

Frequently asked questions

Do ravens remember longer than crows

We would not put it that way. The crow work reports close to seven years for a mask response and we quoted that figure in the crow article. The raven mask study reports four years and stopped because the study stopped, not because the birds forgot. Different labs, different designs, and a four year number is a floor rather than a ceiling.

Will a raven attack me if it decides I am a threat

What was measured in the four year study was scolding, which is alarm calling. The dependent variable was how much of each presentation the bird spent making noise. Nothing in that design tested contact, and we are not going to extrapolate from calling to attacking.

Can I get on a raven's good side by feeding it

The trading study is the closest thing to evidence, and it is only close. Birds preferred a person who completed the deal over a person who cheated. That is captive birds in a structured exchange, not a garden, and the fair partner was not being generous so much as honest.

Will the rest of the group learn about me

In that experiment, no. The two ravens that only watched showed no preference afterward.

Is a raven smarter than a crow

Not a question these studies can answer. They were run in different countries, on different species, using different measurements, for different lengths of time.

References

1. Reference: Rapid Learning and Long-Term Memory for Dangerous Humans in Ravens (Corvus corax) | Blum, Fitch and Bugnyar, Frontiers in Psychology, 2020
2. Reference: Ravens found able to remember people who tricked them up to two months later | Phys.org, reporting the 2017 trading study
3. Reference: Ravens Remember People Who Have Wronged Them, Study Says | Time, on the same study
4. Reference: Common Raven Life History | The Cornell Lab of Ornithology, All About Birds