May 2010


Strong Evidence of Empathy Among Ravens

Reasons for singling out humans as somehow "special" and fundamentally more highly evolved than other animals keep on getting torn down by science. Many examples are out there, even from the past year or two.

Clearly, humans aren't as special as we sometimes think we are.

Recent research by Orlaith Fraser (University of Vienna, Austria) and Thomas Bugnyar (Konrad Lorenz Forschungstelle, Austria) tears down the supposed barriers between humans and other animals even further. They have strong evidence of empathy among ravens.

Basic experimental design.

Testing for empathy among ravens isn't straightforward. You obviously can't just ask a bird what he's thinking, and expect a reply of, "I'm doing fine, buddy, but the meat could have used some mayonnaise."

These scientists have been studying a group of captive ravens for a while now, and recently reported the security, compatibility, and value of all two-way relationships within this group. This enabled them to focus their study on the juvenile ravens.

The juvenile ravens have a wide range of relationships with the other ravens (compared to adults), including valuable relationships other than with a mate. This enabled the scientists to (somewhat) readily collect many data points for their study.

The thirteen juvenile ravens under study were all hand-reared (two of them died from predation over the course of the study), and two adult ravens were also included for a portion of the study. Their home was meant to be somewhat natural, embellished with trees, pools, stones, tree trunks, and branches, and the ravens were given plenty of food and water.

The scientists recorded all instances of aggression; chase-flight or hitting was defined as high-intensity conflict, and forced-retreat was defined as low-intensity. The identities of the aggressor and victim were also recorded.

Defining a conflict/control scenario.

If the scientists watched a raven get into a fight, and then come into contact with another raven, that's not enough to rigorously test for empathy. They needed a more rigorous standard, one that cannot be criticized as biased.

The scientists followed a standard protocol, originally developed nearly thirty years ago for monkeys, for resolving this challenge. After each raven conflict, the victim was watched for 10 minutes; all affiliative actions with and aggressive actions against the victim were recorded.

For a control, the same victim raven was watched at the same time the next day. Observance was postponed if aggressive action was observed 10 minutes before the scheduled control time.

If a raven was a victim, closely interacted with another raven within the next 10 minutes, and closely interacted with another raven within the same 10 minute time window the next day (i.e. interactions expected to happen naturally), this was a usable conflict/control scenario. The time delay between the original aggression and the subsequent interaction with another raven, compared to the time delay the next day, was scrutinized for possible evidence of empathy.

In brief, the ravens were watched to see if a conflict enhanced the probability of an affiliative interaction. The scientists' data is based on 152 conflict/control matchings observed from August 2004 to June 2006, during the day; clearly, a lot of bird-watching was required to compile a useful data set.

Evidence for solicited and unsolicited bystander affiliation.

The scientists observed in total 16 post-conflict affiliations between former opponents after the 152 observed conflicts. However, these affliliations on average occurred statistically faster neither after the conflict nor during the control, suggesting no reconciliation among raven opponents, which would be expected if the benefits of reconciliation are outweighed by the risks of renewed conflict.

Results for unsolicited bystander affiliation (initiated by a bystander, not the victim) were different. Here, 38 ± 6% of the affiliations occurred faster after the conflict, and 15 ± 4% of the affiliations occurred faster during the control.

Therefore, affiliations were 2.5 times more likely to occur quickly after a conflict than would be expected without a conflict. This suggests unsolicited bystander affiliation among ravens.

Results for solicited bystander affiliation (initiated by the victim, not the bystander) turned out similarly. Here, 27 ± 3% of the affiliations occurred faster after the conflict, and 14 ± 4% of the affiliations occurred faster during the control.

Therefore, affiliations were 1.9 times more likely to occur quickly after a conflict than would be expected without a conflict. This suggests solicited bystander affiliation among ravens.

By a similar approach, the scientists found that the victim ravens were no more likely to attack a bystander after losing a conflict, and were more likely to get into another conflict with the aggressor after being attacked. This collectively sets the stage, but is not itself evidence, for empathy; this issue is discussed next.

Evidence for empathy among ravens.

The scientists observed an increasing probability of post-conflict unsolicited bystander affiliations after more intense conflicts (when the victim was more likely to be distressed). They did not observe a similar trend between solicited bystander affiliations and conflict intensity.

Furthermore, the bystander was more likely to share a valuable relationship (a close social bond) with the victim raven than the aggressor. This set of evidence meets a reasonable standard for the demonstration of empathy among ravens.

A note of caution must be introduced. The scientists did not record vocalizations among the ravens.

If vocalizations preceeded affiliative interactions, this would cast doubt on their distinction between solicited and unsolicited affiliative interactions. On the other hand, note how hard it would be to check on this possibility; you'd likely need some sort of recording device on each raven.

Other notes of caution are that this study was performed with a relatively small number of ravens (surely to make the experiments more manageable), with captive rather than wild ravens, and with juveniles rather than adults. However, this study does help us understand how ravens manage their social relationships and share the costs of living together, and provides strong initial evidence of empathy among non-primates. for more information:
Fraser, O. N., & Bugnyar, T. (2010). Do Ravens Show Consolation? Responses to Distressed Others PLoS ONE, 5 (5) DOI: 10.1371/journal.pone.0010605