BBC Earth newsletter
BBC Earth delivered direct to your inbox
Sign up to receive news, updates and exclusives from BBC Earth and related content from BBC Studios by email.
Animals
Do you ever look at your dog, or cat, or a bird flying by the window and wonder: “What exactly are you thinking?”
So many of us treat the animals around us as extensions of ourselves, having long, involved conversations with our furry or feathered friends and assuming they can completely comprehend everything we are saying to them.
But the gulf between human and non-human communication is still vast. But will there ever be a time when, Dr Dolittle style, we’ll be able to talk to animals and they can respond in kind? While it seems to be the stuff of fiction and fantasy, could AI and new technologies further the cause of animal translation?
Animals communicate by a wide variety of different means - from gestures, colour changes, sounds, chemical release, vibration, thermo-reception and touch. The numerous reasons for these varied signals; from attracting mates to providing warnings about predators (or, conversely, alerting others to the location of prey). Most of this communication is a transfer of information between the same species, but inter-species communication is also abundant, especially between a predator and its prey. Think of an insect or fish changing colour to warn off attackers or alert them to their toxicity. But can these phenomena occur between humans and other animals?
Any dog owner knows that you can convey your emotions and have them understood by your pet. If you scold a dog for some wrongdoing, the dog knows it is being told off and responds accordingly. But there is a danger in thinking that human communication is accessible to animals. As the American philosopher Thomas Nagel argued, the consciousness of animals is inaccessible to human beings because of our own consciousness and specific human experience. The rest of the animal kingdom experiences the world in a way completely alien to us. So, even if we could understand what their squeaks, mews and snorts translate as, would we be able to grasp their meaning?
Another problem with attempting to understand the language of animals, if such a thing could even exist, is that we have so little understanding of our own language and how it developed. We don’t really know how or why we developed a system of communication so remote from our animal brethren. One theory is that it developed slowly, starting with gestures, then noises, until a whole vocabulary was established. Another theory proffers the idea that language began more like bird song, a complex series of sounds that eventually had meaning layered upon them. But if either of these cases are true, why didn’t other species evolve and develop speech in the same way? And do any creatures use language in the same way as us; not just to respond to immediate external stimuli but to convey abstract thoughts and invoke creativity?
So how far have we advanced in our attempts to translate the signals of animals into something we can comprehend? Koko, the famous gorilla who mastered around 1000 sign language gestures was viewed critically by certain areas of the scientific establishment, who felt she had little or no understanding of what her gestures actually meant. Elsewhere, parrots can learn and vocalise up to 100 words and express needs and desires, such as wanting food or to be moved. However researchers don’t know if there is any understanding underlying these words. Is it just simple learned behaviour? Scientists have also been able to convey a basic level of communication understanding with dolphins, elephants, orcas, horses, dogs and primates.
So how can developing technologies help in our understanding of animals. For many years now, companies have been claiming they’ve created apps or interfaces that can ‘translate animals’, but many of these have been discredited or can only offer a very basic level of communication (my dog is angry, sad, sleepy). The development of AI systems and deep learning algorithms have advanced the idea of using big data to assemble some kind of animalistic lexicon.
One researcher, who has been studying the sounds of prairie dogs for decades, is convinced that a translator for domestic dogs, again using machine learning, could be with us within a decade. But the same problem remains; how will we ever know that what we think a dog is barking at us is actually what the dog is trying to get across? Will an animal ever be able to tell us that our translation is accurate? And should we even be trying to understand animals in the first place? It’s possible that, in the case of farm animals or even our beloved pets, they might tell us something we don’t want to hear.