Picture this: You’re at a bar and someone clearly intoxicated starts telling your friend their grand theory about how the Titan submersible implosion was faked. Your friend locks eyes with you, clearly wanting to leave this dreadful conversation. She makes eyes to the door. Following someone’s gaze may seem like a simple act, but it has profound implications for the evolution of intelligence. And humans are far from the only animals that do it.

A recent study of bottlenose dolphins in the journal Heliyon adds to previous research identifying the ability to follow the gazes of members of other species — a visual and cognitive trick that may relate to the development of empathy — across a wide range of mammals, not just humans and our fellow primates. What’s even more interesting is to trace this ability through not just the mammal family but beyond, to reptiles and birds — and perhaps back as far as the Jurassic period.

Doing so reveals not just aspects of how the human capacity for empathy may have evolved from traits seen in our ancestors, but also displays the mysterious details of evolution by natural selection. While not driven by any conscious or guiding force, it can in a way be seen as nature’s imagination — which sometimes comes up with the same ideas over and over again.

Putting yourself in another’s shoes

Gaze following can help an animal identify predators or see what tasty treats their same-species competitor has discovered, among other useful things.

To evaluate animals’ abilities to follow the direction a human experimenter is gazing — for example, noticing the experimenter looking at food and then checking back to be sure before going for the reward — researchers teach the animals how to independently gain a reward. Then, scientists being mean buggers, will give them a similar task that is unsolvable: this is called the “impossible task paradigm.”

But, given an impossible task by Elias Garcia-Pelegrin and his team of researchers (who did not respond to an email interview request from Salon), bottlenose dolphins were not, in fact, driven mad in frustration; instead, they demonstrated the ability to use human attentional cues, staying still and quickly alternating their gaze between the experimenter and the object of the impossible task — while giving up the gaze alternation as soon as the lead experimenter’s back was turned towards them.

Of note: gaze following isn’t a single thing; the impossible task literature divides it into various types, which may suggest different cognitive abilities on the part of the experimental animal. “High-level” gaze following, like the dolphins demonstrated, involves putting oneself in the shoes of another by watching where they are looking to see from the other’s perspective.

In general, by identifying important objects in their environment, an animal’s ability to follow the gaze of another, including another species, may form a basis for advanced social cognition, paving the way for cooperation and empathy.

One such high level type, “geometrical gaze following,” occurs if you block the thing that the other is looking at so the subject can’t see it, so that they will physically reposition themself to see what others are seeing. Geometrical gaze following isn’t even seen in human children before eighteen months of age – and yet wolves, apes and monkeys, and birds of the crow (corvid) and starling genuses have all been found to engage in it. You’ll notice, perhaps, that the trait has therefore been seen in various mammal families (primates and the dog-like animals, called canids), as well as some but not all birds. But what does this mean?

Converging on a point

Most likely, it suggests that visual perspective-taking or gaze following evolved independently in mammal groups that had already diverged earlier in their history. For example, experimental evidence suggests it might have arisen at similar times, though separately, in both the monkey ancestors (primates) and dog ancestors (canids) This is called convergent evolution, where evolutionarily distinct groups that occupy similar environmental roles (or “niches”) evolve similar traits.

---

Want more health and science stories in your inbox? Subscribe to Salon's weekly newsletter Lab Notes.

---

“The sort of simple way that I typically define convergent evolution,” Tim Sackton, director of bioinformatics at Harvard University’s FAS Informatics Group, told Salon, “is if there’s a trait that you see in some species, whatever it is, that evolved independently.”

That is, the trait isn’t one that the species you’re comparing got from their common ancestor, but one that emerged in totally different lineages.

“Many other traits seem to be solutions to common problems,” Sackton said. “And so natural selection sort of optimizes for organisms to converge on that same phenotype.”

By phenotype, Sackton means the actual expression of that trait, like having flippers or engaging in gaze following, as opposed to its genotype, meaning the genetic makeup that results in that trait.

Examples of convergent evolution include the similarly streamlined teardrop body shape that evolved in ichthyosaurs, sharks, tuna and dolphins — a response driven by natural selection in similar ocean environments; the camera-like eye structure that evolved independently in vertebrates, including humans, and in cephalopods like squid or octopuses; or certain fish in both the Arctic and Antarctic seas, only very distantly related, which independently evolved antifreeze proteins to protect their tissues and blood from the extreme cold.

Likewise, it seems that gaze following is an aspect of social cognition that has proven its worth as a “solution” to problems for a variety of evolutionarily distant groups.

As a bioinformatician, Sackton’s interest lies in trying to understand what part of the genome of very different evolutionary groups can lead to similar traits being expressed. The traits that strike us as convergent sometimes actually relate to similar proteins being produced by the expression of related genes in these very distant species; sometimes, though, the convergent traits are more superficial than that and only seem similar without having an underlying genetic basis in common.

Take the convergent evolution of flippers. Sackton and colleagues have found that areas of the genome that regulate the development of the hindlimbs are at play in the very divergent types of animals in whom hindlimbs devolved into flippers. By contrast, Sackton’s collaborator Nathan Clark has found that in the loss of eyesight that occurs sometimes in the evolution of many unrelated subterranean animals, the genome changes from that of their non-subterranean ancestors in similar ways to do with genes coding for proteins expressed in the lens, cornea or other parts of the eye. Whether the genes in question relate to the developmental process or to the expression of proteins, Sackton and Clark write that we’re finding that there’s often a lot more genetic convergence — similar things going on at the level of genes — underpinning the similarities we see between unrelated organisms than you’d expect.

So far, there doesn’t seem to have been much research into the genetic underpinnings of gaze following in animals — although there has been some looking at humans, in whom impaired gaze following can be a sign of conditions such as autism spectrum disorder.

Diverging again

What about birds and their reptilian relatives? Why would some have advanced gaze following abilities and some not? A study published last year in Science Advances looks at Archosaurs, the group that includes birds, crocodilians and their dinosaur ancestors, providing some evidence about this.

Researchers Claudia Zeiträg, Stephan A. Reber, and Mathias Osvath compared paleognaths, the most neurocognitively “basic” of birds, with crocodilians, birds’ closest living relatives. They found that the alligator, a crocodilian, was unable to really grasp advanced visual perspective taking. However, both the paleognaths (those birds most similar to their earliest bird ancestor, such as the kiwi, the ostrich and the cassowary) and non-paleognath birds (more specialized birds — a nice duck, say, or a swallow — that have evolved characteristics that make them less similar to the earliest bird ancestors) all engaged in gaze following. They even exhibited checking-back behavior at the level of apes.

Alligators do follow gazes into the distance, but this simpler form of gaze-following is a feature shared by all amniotes (that is, all of the four-legged animals plus descendants of four-legged vertebrates, like birds).

The visual perspective-taking exemplified by geometric gaze following, write Zeiträg and her colleagues, “is a form of functional representation, leading to behaviors that correspond to the fact that the other has a different perspective and that its gaze refers to an object.” Even those basic birds – in scientific terms, “neurocognitively most conserved” – showed both geometric gaze following and the ability to check back, and that “presupposes the expectation that the other’s gaze is directed at something, which cannot currently be seen. Checking-back is a behavior signifying such an expectation,” as they put it.

In human children, checking back precedes gaze following, and children show evidence of it by about eight months of age. On the other hand, among birds, the more advanced geometric gaze-following has only been observed in some species, but not only the most conserved or "basic" of them. This might mean a particular species evolved to lose this trait, or that we simply haven’t looked hard enough for its presence in different bird species.

Similarly, while among the primates, checking back has only been reported in apes and old world monkeys, there haven’t been very many studies of this in primates, and while one rare such study concluded that new world monkeys — spider monkeys and capuchins — don’t check back, in fact an individual spider monkey was observed checking back in that study, over and over.

This could be a case where “absence of evidence doesn’t equal evidence of absence” of this trait that, if found, would suggest some pretty advanced social and cognitive abilities.

Built for the job… But up for the task?

As well as seeking experimental, observational and genomic evidence of gaze following and visual perspective-taking, a complementary approach is to look at the physical equipment making such abilities possible: that is to say, the eyes, body and brain.

Alligators and crocodiles have eyes that are adapted for seeing in air, not water. Their eyes, placed on either side of their head, give them a wide field of view and scary-good peripheral vision. Their ability to adapt to scan the shoreline without moving their heads makes crocodiles, as one headline about a study on the subject put it, “fine-tuned for lurking”. The kind of low-level gaze-following they engage in is mediated by subcortical structures of the brain–those more “primitive” parts also found in mammals and fish.

Dolphins can use binocular or monocular vision but typically use monocular, giving them a whopping two hundred degree vista from each eye compared to primates’ limited field of view, using our two forward-facing eyes, of around ninety degrees to each side of the midline, sixty below the point of focus, and fifty above. The dolphins thus don’t need to move their heads as most non-primate mammals must if they want to get a good field of sight — a good thing, because their fused cervical vertebrae make that tricky to do.

Basically, where head position and forward eyes is thought to be important for the development of gaze following, in dolphins which use echolocation to recognize objects, it may have evolved in a different way. (Like the dolphins, penguins and ibis, which also have eyes on separate sides of their head, have already been found to show conspecific gaze following.)

In the study of Archosaurs, small birds simply had a harder time actually carrying out visual perspective-taking than big birds, like the rhea or the emu: they weren’t tall enough to see what the experimenter was looking at. As a short person, this author can only sympathize.

Looking at which living species show evidence of advanced gaze following and which don’t suggests that even the more advanced type, and the ability to check for visual references, evolved back in the time of dinosaurs. This also likely means that some dinosaurs evolved the neurocognitive equipment to make these things possible, and that when we start looking into the genomes of these different groups, we’ll find genetic evidence of exactly how these traits are being controlled and whether the dolphin’s gaze following abilities, for example, occur in a similar way to those of the swallow or its Archosaur dinosaur ancestor.

But that doesn’t mean that all dinosaurs exhibited this form of social cognition. Instead, it evolved in some dinosaurs only, probably some time after the Archosaur group, a group that includes both reptiles and birds, divided. This division of the constantly branching evolutionary tree gave rise to the ancestors of today’s crocodiles and alligators in one group, and to the ancestors of bird-like dinosaurs and today’s birds in the other. Tracking convergent evolution through the evolutionary tree is best done with a combination of high-throughput genomic analysis and work that looks at actual animals, whether in museums or in the field, to see how traits are expressed.

As genomic analysis becomes cheaper and easier to do (and as extinction takes a brutal toll on existing species), it can be harder to get funding agencies to invest in studying an animal in the wild – studying its phenotype, or how it expresses traits – than to sequence the DNA of hundreds of thousands of individuals.

“Phenotypic resources are often more challenging,” Sackton told Salon. He stressed the need for collaboration in his work with molecular and organismal biologists to understand how an organism’s ecology might shape what he sees in its genes, and conversely to understand the relevance of the genomic sequencing he does to its phenotype, the traits we can actually observe, like physiology or behavior.

“There’s so many weird things that animals and plants do,” he said. In an alternative pre-history, we might imagine those early gaze-following dinos continuing to evolve, unmolested by giant asteroids that blotted out the sun. Instead of evolution ultimately producing as a dinosaur descendant the clever jackdaw that can follow your gaze to steal your food, we might have a society of empathetic dinosaurs whose early capacity to put themselves in other dinos’ shoes (so to speak) could have led to a complex social world, one in which knowing your dinosaur friend is planning their escape from the dinosaur bar is of great interest.

Perhaps in that alternate world a dinosaur is writing up a story about convergent evolution and the experiments being done to better grasp the amazing, gaze-following abilities of those curious creatures, the bipedal, big-brained, highly social Homo genus of primates and their previously unsuspected empathetic abilities – almost like dinosaurs themselves.