New research complicates the story of dog domestication

A new study published in Animal Cognition suggests that while dogs appear more affectionate and submissive than wolves during greetings with humans, these differences may not be solely due to domestication. Human behavior toward the animals also varied, hinting at possible unconscious biases.

Dogs have lived alongside humans for thousands of years, and many traits we associate with them — such as tail wagging, sustained eye contact, and expressive faces — are often assumed to have evolved as a result of domestication. This has led scientists to question what exactly drives this unique social bond.

One commonly proposed idea is that dogs are “hypersocial,” unusually motivated to seek human attention. Another view holds that dogs are more deferential, tending to comply with humans perceived as dominant.

However, previous research shows that wolves raised in close contact with humans can also engage in cooperative tasks and social interactions with people. This has prompted scientists to question whether dogs’ human-oriented behaviors emerged during the domestication process or are instead rooted in traits already present in wolves.

“We know dogs were domesticated roughly 15,000 years ago but there is still a lot of discussion about what happened during this process – what were dogs selected for to become what we know as dogs today? Previous studies showed that dogs exhibit more human-directed behavior than wolves, but it’s unclear whether this is motivated by hypersociality or their acceptance of humans as dominant figures,” explained study author Svenja Capitain of the Domestication Lab at the University of Veterinary Medicine in Vienna.

“We therefore wanted to compare the facial expressions of human-socialized dogs and wolves during a friendly interaction with a human to try and deduce what the motivation might be – and understand differences in their human-directed facial expressions in the first place, which have not yet been compared. Since the interacting humans couldn’t be blinded to whether they were engaging with a dog or wolf, we also coded for potential biases in human facial behavior as a confounding factor.”

The researchers compared the facial expressions and greeting behaviors of human-socialized wolves and dogs in a highly familiar context. They also investigated how humans responded to the two species, with a particular focus on facial expressions, which are known to convey subtle emotional cues. This bidirectional approach allowed them to examine not only how animals act toward humans, but also how humans might influence those behaviors.

The study took place at the Core Facility Wolf Science Center of the Vetmed University in Austria, where both dogs and wolves are hand-raised from a young age and live in conspecific packs with regular human interaction. Eleven dogs and eleven wolves, all familiar with the greeting routine, participated in the study. An additional group of seven pet dogs with upright ears was included to control for the effects of ear shape, which can affect how facial expressions are displayed and interpreted.

Each animal was greeted by two different human partners, one with whom they shared a strong bond (a primary trainer or caretaker) and another with whom they were familiar but less closely connected (a researcher or student who interacted with them occasionally). The greeting occurred through a fence, mimicking daily interactions at the facility. Human participants wore chest-mounted cameras to record their facial expressions, while multiple stationary cameras captured the animals’ behavior from different angles.

Trained coders used a standardized system called the Dog Facial Action Coding System (DogFACS) to identify specific muscle movements in the animals’ faces, such as inner brow raising or ear rotations. A similar system was used to code human facial expressions associated with positive or negative emotions. In addition to facial movements, the researchers recorded other behaviors commonly interpreted as social or emotional signals, including tail wagging, gaze direction, whining, and physical proximity.

Overall, dogs spent more time near the humans, wagged their tails more, and maintained eye contact more frequently than wolves. Both species showed more of these behaviors toward their bonded person compared to the familiar one, suggesting that relationship strength played a role in their responses. Dogs also exhibited more displacement behaviors like whining and rubbing against the fence, especially when interacting with a bonded person. Wolves did not show this pattern.

When it came to facial expressions, dogs were more likely than wolves to raise their inner eyebrows and slightly more likely to raise their upper lips. “Our findings align with prior studies showing that dogs seek proximity to humans, gaze at them more, and raise their inner eyebrows more than wolves,” Capitain said.

However, these differences were modest and their emotional meanings remain unclear. Other facial expressions, including those associated with stress or positive anticipation, did not differ significantly between species. “We expected more pronounced differences in dogs’ and wolves’ facial expressions given dogs’ long domestication history,” Capitain told PsyPost. “It’s possible that dogs rely more on other body signals, like ear or tail movements, to communicate with humans instead.”

Ear movements revealed more pronounced differences. Wolves were more likely to hold their ears forward, while dogs more often rotated or lowered their ears, which are often interpreted as signs of submission, uncertainty, or appeasement. These differences were not explained by ear shape, as the pet dogs with upright ears showed similar patterns to the dogs with floppy ears.

On the human side, facial expressions varied significantly based on the species of the animal they were greeting. Humans displayed more intense, more frequent, and more positively valenced expressions when interacting with dogs than with wolves, regardless of whether the person had a close relationship with the animal. This pattern held even among highly experienced handlers who had worked with both species for many years and were instructed to treat them equally.

“We were also surprised to find such pronounced biases in the human facial expressions, even among those with extensive experience working with both species,” Capitain explained. “This raises questions about whether humans mirrored the animals’ behavior (e.g., responding to dogs’ intense approaches with more expressive faces) or whether the human influenced the animals’ behavior.”

The results suggest that humans may harbor implicit biases toward dogs, responding to them with more warmth and expressiveness than to wolves even in controlled, familiar settings. These subtle differences in human behavior may, in turn, influence how the animals behave, either in the moment or over the course of their development.

“Interestingly, there were no major differences in the species’ human-directed facial expressions (except for the inner eyebrow raising, which had already been known),” Capitain said. “However, dogs’ human proximity-seeking was accompanied by ear postures generally seen in ambiguous and submissive situations, while wolves showed more forward directed ears, which are associated with confidence and attention.”

“Yet, we cannot interpret these outcomes in terms of species differences or domestication, because we also found marked differences in the human faces. This underlines how complex the human-canine relationship and the study of it can be, and suggests we need to understand our own biases better before drawing conclusions about the animals.”

As with all research, there are some limitations to consider. The sample size was relatively small, which is often the case in studies involving hand-raised wolves. While the researchers used rigorous methods and controls, it remains difficult to determine causality in such a dynamic, bidirectional interaction. It is unclear whether the animals were reacting to differences in human behavior or if humans were unconsciously responding to the animals’ species-typical expressions.

“The human component of the study was exploratory, so increasing the sample size here is a feasible next step,” Capitain noted. “So far, the initial question, what really drives dogs’ human-directed behavior, remains unanswered. But perhaps more importantly, the new questions this research raised is why the humans showed these biases, who was really influencing whom, and how we can guard against potential biases in human-animal interaction research in the future.”

Future studies could help disentangle the direction of influence by using time-series analysis or by standardizing human behavior across interactions. The researchers also hope to explore the origins of the observed human biases and to investigate whether such biases can be measured through implicit association tests or physiological responses.

“We plan to conduct more fine-grained, controlled comparisons of dogs’ and wolves’ human-directed behavior to better understand the domestication process,” Capitain said. “On the human side, we aim to investigate how the facial biases relate to internal biases, where they may originate from, and whether and how they influence the animals. These efforts will not only advance our understanding of dogs and wolves but also improve human-animal coexistence and the rigor of comparative research as a whole.”

“Our lab is dedicated to unravelling the complexities of domestication and human-animal relationships. For more information, you can visit our website (https://domesticationlab.wordpress.com), follow our Instagram channel (@DomLab_Vienna), or explore Friederike’s and Sarah’s book, <em>Wolves and Dogs: Between Myth and Science. We hope our work inspires others to think critically about the fascinating bond between humans and animals.”

The study, “Differences in dogs’ and wolves’ human-directed greeting behaviour: facial expressions, body language, and the problem of human biases,” was authored by Svenja Capitain, Gwendolyn Wirobski, Çağla Önsal, Giulia Pedretti, Valeria Bevilacqua, Sarah Marshall-Pescini, and Friederike Range.