Mirror, Mirror

1471836761_11edeb5212_mHumans use mirrors so easily that we often don’t think about the cognitive processes required to do so (although we do derive entertainment from those animals that just don’t “get” mirrors). Being able to recognize oneself in the mirror doesn’t seem like an evolutionarily advantageous skill, but scientists think that this ability may indicate the possession of other skills more relevant to survival.

For example, mirror self-recognition may be indicative of self-consciousness and, by extension, theory of mind (the ability to think about what others may be thinking); being self-conscious, or having knowledge of the self, may naturally lead to having knowledge of others. These skills are especially important for social animals, as theory of mind allows us to take the perspective of others, which is the basis for empathy. (Evidence for the connection between mirror self-recognition and theory of mind comes from studies with young children, which have found that mirror self-recognition and perspective-taking abilities develop at around the same time.)

But how can we tell whether animals and pre-linguistic children connect what they see in the mirror to their physical selves? The most commonly used measure is the Mirror Test (also called the Mark Test or the Rouge Test), which was developed by Gordon Gallup, Jr. in 1970. After a familiarization period with a mirror, a colored mark is placed on an area of an animal that can’t be seen without the use of a mirror (usually somewhere on the face). The animal is then exposed to a mirror again, and her behavior is closely monitored. If the animal touches the mark on her own face (rather than the mirror), then she has “passed” the Mirror Test and demonstrates mirror self-recognition.


Not quite there yet…

When first exposed to a mirror, most animals (including humans) exhibit social behaviors like lip smacking and attempts to play, indicating that they perceive their reflection as a conspecific. After some cognitive development and experience with mirrors, however, some animals will demonstrate mirror self-recognition (for humans, this occurs around 18-24 months of age).

Unsurprisingly, humans’ closest evolutionary relatives, the great apes, demonstrate mirror self-recognition (the great apes include chimpanzees, bonobos, gorillas, and orangutans; all have passed the Mirror Test). Additionally, some very distant relatives of humans, but who are also highly social and have demonstrated advanced cognitive abilities, have passed the Mirror Test: dolphins and elephants.

One unexpected species that has passed the Mirror Test is the magpie. But when you consider their other cognitive abilities, it’s not so surprising: magpies have also demonstrated the abilities of tool use, perspective-taking, and foresight. (Why would these abilities be particularly helpful to magpies? Researchers theorize that they enable magpies’ prolific thievery.)

7193567450_b7ebd10bb2_mInterestingly, lesser apes (gibbons) and monkeys fail the Mirror Test, suggesting that mirror self-recognition and all the attendant cognitive abilities (self-consciousness, perspective-taking, etc.) evolved after the evolutionary split between great and lesser apes (which occurred after the split between apes and monkeys). Elephants, dolphins, and magpies, then, must have evolved the abilities through convergent evolution.

So it seems pretty simple: pass the Mirror Test, and you demonstrate mirror self-recognition and its associated cognitive abilities. Yet, as with most matters in animal cognition, this one is far from cut-and-dry.

Some researchers have argued that the Mirror Test isn’t appropriate for many animals. First, a colored mark on the face may not be salient enough for some animals (it may not stand out or be important enough for the animals to notice). Additionally, just because an animal doesn’t touch or try to remove the mark doesn’t mean he doesn’t recognize his reflection; perhaps he notices the mark but just doesn’t care. Finally, what about animals who have poor vision, or who rely primarily on other senses? Surely we can’t say they lack the abilities of self-consciousness and perspective-taking simply because they don’t pass the Mirror Test.

This issue is complicated even when studying children: some studies have shown vast cultural differences in performance on the Mirror Test. For example, many of the Kenyan children in one study froze in response to seeing their reflections. While some of these cultural differences could be attributed to differences in experience with mirrors, researchers think they are likely more related to different parenting styles and differences in how the children understand the task. Children raised in cultures with a high emphasis on obedience, for example, may recognize their reflections but be unsure of whether they’re allowed to investigate or remove the mark.

1921632741_baee2c47b8_mThese criticisms of the Mirror Test have prompted some researchers to try to find other methods of gauging whether animals demonstrate mirror self-recognition. Some possible alternative indicators include mirror-guided self-directed behaviors (like using the mirror to examine one’s body) and the disappearance of social responses to the mirror. Some researchers attempted to use these indicators as proof that macaques can recognize themselves in the mirror, although their interpretation has been questioned. (I highly recommend reading these papers and judging for yourself – you can find the original paper here. Unfortunately, the rebuttal paper isn’t available for free, but the same authors briefly discuss their criticisms in this article.)

This brings me to a tangential point about scientific research in general. These kinds of exchanges regularly occur in science, regardless of the particular field. Often, one researcher (or a group of researchers) finds an issue with another researcher’s methods or conclusions. Sometimes she will just write a rebuttal pointing out the flaws she sees or offering an alternate interpretation of the results, but she may also conduct her own experiment, fixing any methodological flaws from the original study. While it is undoubtedly frustrating to have your methods and conclusions questioned, it ultimately leads to better science by pushing researchers to really think about and improve task design and to be very careful about interpreting results.

It is also the theoretical basis of how we disseminate findings in the scientific community: after careful consideration by other scientists in the field (peer review), we publish not only the conclusions of our experiments, but also our exact methods, our data, and the related prior research that helps lead us to those conclusions. In essence, we tell a story about our research. Others can then follow along and, rather than taking our word for it, decide for themselves whether our story is compelling and what makes it so (or not).

(As I said, though, this is all in theory. There is currently quite a bit of debate concerning how papers are chosen for journals, whether access to these papers should be free to everyone, and a host of other issues relating to publishing scientific research.)

174px-Mirror_test_with_a_BaboonBut I digress. Hopefully this was a thought-provoking summary of what mirror self-recognition can (but possibly can’t) tell us about certain cognitive abilities. I’ll leave you with this video about the Mirror Test in primates (starts at 1:00), and this one about dolphins and elephants.





Anderson, James R., and Gordon G. Gallup. “Do rhesus monkeys recognize themselves in mirrors?.” American journal of primatology 73.7 (2011): 603-606.

Anderson, James R., and Gordon G. Gallup Jr. “Which primates recognize themselves in mirrors?.” PLoS biology 9.3 (2011).

Broesch, Tanya, et al. “Cultural variations in children’s mirror self-recognition.” Journal of Cross-Cultural Psychology 42.6 (2011): 1018-1029.

de Waal, Frans BM. “The thief in the mirror.” PLoS biology 6.8 (2008).

Gallup, Gordon G. “Chimpanzees: self-recognition.” Science (1970).

Plotnik, Joshua M., Frans BM De Waal, and Diana Reiss. “Self-recognition in an Asian elephant.” Proceedings of the National Academy of Sciences 103.45 (2006): 17053-17057.

Prior, Helmut, Ariane Schwarz, and Onur Güntürkün. “Mirror-induced behavior in the magpie (Pica pica): evidence of self-recognition.” PLoS biology 6.8 (2008): e202.

Rajala, Abigail Z., et al. “Rhesus monkeys (Macaca mulatta) do recognize themselves in the mirror: implications for the evolution of self-recognition.” PLoS One 5.9 (2010): e12865.

Reiss, Diana, and Lori Marino. “Mirror self-recognition in the bottlenose dolphin: A case of cognitive convergence.” Proceedings of the National Academy of Sciences 98.10 (2001): 5937-5942.

Suddendorf, Thomas, and David L. Butler. “The nature of visual self-recognition.” Trends in cognitive sciences 17.3 (2013): 121-127.


Also interesting:

Platek, Steven M., and Sarah L. Levin. “Monkeys, mirrors, mark tests and minds.” Trends in ecology & evolution 19.8 (2004): 406-407.

Suddendorf, Thomas, and David L. Butler. “Response to Gallup et al.: are rich interpretations of visual self-recognition a bit too rich?.” Trends in cognitive sciences (2013).


Birds of a Feather

I’ve talked a lot about monkey minds on this blog, but it’s about time I got to the bird brains. Birds may have a reputation for being stupid (hence the disparaging term “birdbrain”), but they actually have some pretty incredible cognitive abilities. In fact, pigeons are one of the most widely studied animals in animal cognition labs.

320px-NZ_North_Island_Robin-2This week, though, I’m going to discuss a study that looks at cognition in a different bird, the North Island Robin of New Zealand, and in their natural habitat rather than in a lab. In this study, Barnett et al. (2013) investigated whether these birds could discriminate between familiar and novel humans.


Research has shown that birds can recognize humans who have previously approached their nests or captured them for tagging, which makes evolutionary sense: being able to recognize predators allows the birds to respond appropriately (e.g. fly away), increasing their chances for survival.

In this experiment, though, the researchers were not acting like a threat; instead of approaching the bird, a researcher placed a mealworm on the ground, then stood one meter away and timed how long it took for the bird to eat the mealworm. This encounter is likely much less stressful to the bird than if the researcher approached the bird’s nest or tried to capture it. This could actually affect the bird’s memory of the researcher because of something called the corticosterone response. When a non-human animal is stressed, its body releases corticosterone (cortisol is the human equivalent). One effect of corticosterone is to improve the formation of memories, so birds may be more likely to recognize a familiar human they’ve previously encountered in a stressful situation than one encountered in a less stressful or neutral situation.

In order to test whether the robins could discriminate between familiar and novel humans, Barnett et al. used a habituation task: they repeated the above procedure (timing how long it took for a robin to eat the mealworm) once a day for 7 days, always with the same researcher. On the 8th day, they used a different researcher.

Animals tend to be cautious when exploring novel things, so the researchers expected that the robins would be slower to eat the mealworms on the first few days. Eventually, though, they would habituate to the researcher and their “attack latency” (how long it took to eat the mealworm) would decrease. The critical data, then, is the robins’ attack latencies on the 8th day, with the new researcher. Longer attack latencies on Day 8 compared to Day 7 would indicate that the robins perceive the researcher as novel, showing that the robins can discriminate between familiar and novel humans. However, no change in attack latency on Day 8 would suggest that the robins cannot discriminate between familiar and novel humans.

But Barnett et al. were interested in more than the overall discriminative ability of North Island Robins; they also wanted to see how individual differences in behavior from bird to bird were related to this discriminative ability. Animals, like humans (although probably to a lesser extent), exhibit variations in behaviors and responses on an individual basis. The human equivalent of this is personality, although researchers also refer to it as temperament. This topic is somewhat controversial, but there are two generally accepted requirements for a behavioral trait to be considered a personality trait. First, the trait must be stable and consistent (not just a one-time behavior). Second, that trait must be related to other traits. For example, boldness, a personality trait indicating risk-taking tendency, is also related to exploratory behavior and aggressiveness.

159px-Petroica_longipes_-_Adam_Mark_Lenny_01Previous research indicates that differences in personality traits are associated with differences in learning and interacting with the environment, which is why Barnett et al. wanted to investigate individual behavioral differences in the robins. Based on the attack latency data from Day 4 (when all the animals had habituated) to Day 7, the researchers split the robins into two “behavioral types”: fast attackers and slow attackers.

The researchers found that the fast attackers did not have an increased attack latency on Day 8, suggesting that they could not discriminate between the familiar and novel researcher (or that they could discriminate between them, but didn’t perceive the novel researcher as threatening). The slow attackers, on the other hand, did have an increased attack latency on Day 8, showing that they could discriminate between the researchers. Moreover, when Barnett et al. compared the attack latencies between the two groups over all 8 days (not just Day 4 through Day 8), they found that the fast attackers habituated to the researcher on the second day, while the slow attackers didn’t habituate until the fourth day. This result agrees with previous findings that bolder animals are quicker to explore novel environments and form routines.

Barnett et al. offer a couple hypotheses of mechanisms behind these behavioral differences. Perhaps the slow learners paid more attention to their environment during the habituation phase (Days 1-7), allowing them to better perceive when the researcher was different. Another theory is that the slow attackers may have a greater corticosterone response than the fast attackers. This could cause the slow attackers to have a much better memory for the familiar researcher.

In addition to showing that North Island Robins can discriminate between familiar and novel humans, this study demonstrates that personality traits can affect individual animals’ behavioral responses. It also suggests that animal cognition researchers should take into account the personality traits of individuals when conducting cognition experiments.


Source Cited:

Barnett, Craig, et al. “The ability of North Island robins to discriminate between humans is related to their behavioural type.” PloS one 8.5 (2013): e64487.

Paying it Forward

8392492488_b3baaee68a_mYou’ve probably heard of the concept of paying it forward: someone does a small act of kindness for you and, instead of repaying that person, you pay their kindness forward by doing an act of kindness for someone else. This may not seem evolutionarily advantageous at first (the smart thing to do would be to simply accept the act of kindness and not expend energy or resources paying it forward), but remember that we are social animals. The prosocial behavior of paying it forward is beneficial to maintaining the close social ties that enable our species to survive.

So it makes sense that we might see other social animals, like non-human primates, pay it forward as well. However, many scientists think that a pay-it-forward mentality requires some higher cognitive abilities that non-human primates just don’t possess. One of these is the ability to feel and understand gratitude, which hasn’t been found in non-human primates. Social and cultural norms likely also play an important role in paying it forward. For example, if someone does something nice for you, and you don’t either do something nice back or pass it on, your reputation might suffer.

On the other hand, some scientists argue that these higher cognitive abilities aren’t required for pay-it-forward behaviors. Rather, animals could simply use generalized reciprocity (“help anyone, if helped by someone”). Generalized reciprocity is simple; it doesn’t require gratitude, or concern for one’s reputation, or taking the perspective of others, or inhibiting the impulse to look out only for oneself. So through the mechanism of generalized reciprocity, social animals without the higher cognitive abilities of adult humans can still exhibit pay-it-forward behaviors. (Why do I say “adult” humans? Because children don’t initially possess these higher cognitive abilities – they must develop them.)

160px-Cebus_apella_eating_grapes-0004Some researchers investigated whether non-human primates and human children pay it forward. Leimgruber et al. (2014) had capuchins and 4-year-old children play a game where an “actor” could choose between equal rewards for her and a “recipient”, or unequal rewards. In both cases, the reward for the actor was the same (a grape for the capuchins and 4 stickers for the children). In the equal option, the recipient received the same reward as the actor. In the unequal option, the recipient received less (spinach for the capuchins and 1 sticker for the children).

In order to see whether the capuchins and children would pay it forward, each trial of the game consisted of two rounds. In the first round, Capuchin (or Child) A would be the actor and choose the reward for herself and the recipient (Capuchin B). In the second round, Capuchin A would leave, and Capuchin B would become the new actor, who would then choose the reward for himself and the new recipient (Capuchin C).

Leimgruber et al. were interested in the choice of Capuchin B in the second round. Would he be more likely to choose the equal reward for Capuchin C if Capuchin A had chosen the equal reward for him in the previous round? Such a result would show that capuchins could indeed pay it forward.

320px-Brown_capuchin_(7958443592)The researchers found that, when the equal reward had been chosen for them in the first round, capuchins and children chose the equal reward in round two significantly more often than chance (80% and 70% of the time, respectively). Interestingly, the researchers also found that both capuchins and children also “paid forward” unequal rewards. When the unequal reward had been chosen for them in the first round, capuchins and children chose the unequal reward in round two 75% and 72% of the time, respectively.

Together, these results demonstrate that capuchins and children pass on both positive and negative outcomes (i.e. equal and unequal rewards). This suggests that, instead of a “help-if-helped” mechanism, generalized reciprocity may be more like “give-what-you-get”, where the “give” and “get” can be positive OR negative.

159px-Cebus_apella_01Leimgruber et al. suggest that the results could also be due to affective processes. “Affect” refers to basic positive and negative feelings, which have been found in many species. So, in the case of this study, receiving an equal reward could put an individual in a positive affective state, which could in turn make that individual more likely to give an equal reward in the next round. (The emotions we talk about having as humans, such as shame and the aforementioned gratitude, are considered to be secondary emotions that require higher and more complex cognitive abilities that most non-human animals don’t have.)

So does this mean that when a stranger buys my morning coffee and I pay it forward by helping my neighbor shovel his driveway, I’m influenced by “give-as-you-get” or a positive affective state? Probably not, say the researchers. Adult humans do have concern for their social reputations, can feel gratitude, and can take the perspectives of others, and we are likely greatly influenced by these considerations. But the simple “give-as-you-get” mechanism and the effect of affect form the base upon which these considerations build, allowing humans to make complex social decisions and have incredibly rich social relationships.

Sources Cited:

Leimgruber, Kristin L., et al. “Give What You Get: Capuchin Monkeys (Cebus apella) and 4-Year-Old Children Pay Forward Positive and Negative Outcomes to Conspecifics.” PLOS ONE 9.1 (2014): e87035.

van Doorn, Gerrit Sander, and Michael Taborsky. “The evolution of generalized reciprocity on social interaction networks.” Evolution 66.3 (2012): 651-664.