After spending at least 30,000 years living alongside humans, dogs have become really good at reading our emotions.

Not only can they understand some of our most iconic words, signs and gestures, studies have shown these animals can recognise our facial expressions and determine whether our responses are positive or negative.

Even our closest relative, the chimpanzee, doesn't understand us to the same extent. But how well do we actually know our dogs?

New research suggests that primarily depends on where and how you grew up, not on whether you've ever actually owned one.

While most studies have looked into our pets' understanding of us, surprisingly few have turned the focus around - although, there are a few contending theories.

Under the co-domestication hypothesis, for instance, the close evolution of both humans and dogs is said to have resulted in an inherent reciprocal understanding. As both species grew closer over time, it follows that dogs evolved the ability to read human emotions and vice versa.

According to this theory, even though dog owners may be better at reading their pets than others, in the end, this ability is partially present in all humans.

It's an interesting idea, but actual research looking into it has provided mixed results. Some studies have shown those people who are inexperienced with dogs are actually better at reading canine emotions, while others have found no difference whatsoever between dog-owners and non-owners.

A new, comprehensive study has now brought up partial support for this hypothesis. While some dog emotions like anger and happiness were easily recognised from a young age - suggesting they could be inherent - the authors argue this ability is mainly acquired through experience.

In other words, recognising a dog's facial expression is not an evolutionarily selected trait, but rather a consequence of our cultural environment.

"In particular, participants with more general dog experience were overall more proficient at recognising dog emotions than participants with less general dog experience," the authors conclude.

The study was conducted on 89 adults and 77 children, between ages five and six, who were either Muslim European or non-Muslim European. Each participant was categorised depending on their culture's attitude toward dogs, as well as their own personal history of dog ownership.

They were then shown a series of photographs, including dogs, chimps and humans, and asked to rate their happiness, sadness, anger or fear.

While some dog emotions were easily recognised early on in life, the findings suggest this skill is mostly acquired through age and experience, even if that experience isn't firsthand.

Children in the study, for instance, were quite limited in their recognition of dog emotions, regardless of their history with canines or their cultural upbringing. Whereas for adults, experience really mattered.

The authors found that those participants who grew up in Europe's dog-positive culture - where canines are prioritised and closely integrated into society - were generally better at recognising dog emotions, even if they never owned a dog themselves.

"These results are noteworthy," says evolutionary anthropologist Federica Amici of the Max Planck Institute, "because they suggest that it is not necessarily direct experience with dogs that affects humans' ability to recognise their emotions, but rather the cultural milieu in which humans develop."

The study is relatively small in its scope and does come with several limitations. All the dogs used in the study, for instance, had the same German shepherd-like face; children under five were not tested; only two different cultural groups were examined and participants were asked about their experience with dogs before the trial, which could have influenced their answers.

Nevertheless, the preliminary findings suggest that reading dog emotions does not appear spontaneously in young children, but rather is acquired over time from the society we live in.

The only exception were the emotions of anger and perhaps happiness, which even children could recognise in a dog's face. 

"These results seem to support the co-domestication hypothesis, in that even children with minimal experience correctly interpret some dog emotions," the authors write.

"The ability to recognise anger is clearly adaptive, as it provides immediate fitness benefits by conveying crucial information about possibly dangerous situations, and thus bears higher survival costs."

There is, however, another explanation. Recognising anger is quickly learned through experience and it could be that this skill was acquired before age five.

Nevertheless, the authors argue, even young Muslim children in Morocco, where dogs are not prioritised as pets, can successfully recognise canine anger. This suggests, at least for anger, that specific experiences are not nearly as important as specially adapted mechanisms we've evolved over time.

Whereas for other dog emotions, like fear and sadness, recognition does not appear to be innate.

More studies will be needed before we can say for sure, but the results certainly suggest that picking up on subtle canine cues is not something we've acquired through the millennia.

"We think it would be valuable to conduct future studies that seek to determine exactly which cultural aspects affect one's ability to read dog emotions, and to include real-life stimuli and body expressions in addition to instructed stimuli and facial expressions," says comparative psychologist Juliane Bräuer.

"In this way, we could develop a better understanding of inter-cultural variation in emotion recognition." 

The study was published in Scientific Reports.