A tomb in Syria is the source of what may be the oldest alphabetic writing in human history. The unearthing of clay cylinders by John Hopkins researchers appears to predate known alphabetic writing by 500 years dating back to 2400 BCE.
The invention of alphabetic writing, like the Neolithic agricultural revolution, defines the roots of modern human civilization. The clay cylinders appear to have a string of symbols that look like labels to identify the contents of the tomb. We don’t know what these messages mean but in the future, we may find the equivalent of a Rosetta stone to help us decipher the individual symbols themselves and what those who left them behind were describing.
There is no resemblance to the early hieroglyphic symbology invented in Egypt circa 1900 BCE. Symbolic representation of language developed well after humans invented language. We don’t when our species began shaping sounds to have specific meanings because we cannot find any trace in the fossil record. We can look at skeletal remains and see how changes occurred in areas where vocalization occurs, but that’s about it.
Humans aren’t the only species to develop communication skills and language. We know whales and dolphins talk to each other and as we have recently discovered, so do elephants. Birds have been using sound to communicate well before our species came upon the scene. Since they are descendants of the dinosaurs it is likely the latter had similar capacity. Even before animal species communicated through sound, they found ways to express themselves to their fellow members through movement, colouration or altered physical characteristics. It may not be conversation but it certainly is communication.
Communicating With Alien Intelligences
In the 1977 movie, Close Encounters of the Third Kind, humans are visited by otherworld aliens that find a common language through music. In the 2016 movie, Arrival, a linguist is hired to communicate with an alien species that has landed on Earth at twelve locations. The aliens are heptapods, seven-limbed cephalopod-like creatures, whose language symbology resembles circular ink blots and whose speech hardly resembles anything here on Earth. The linguist’s efforts to construct a dictionary present a challenging task with nuances in meaning and repeated misinterpretations of the message the aliens are trying to deliver seen as either peaceful or threatening.
In neither movie did humans have artificial intelligence (AI) agents to help decipher meaning from the communications and symbology produced by these alien visitors. Machine learning, however, and the development of neural networks and natural language processing may turn out to be how we eventually discern messages sent to us from intelligent extraterrestrials.
Language And AI
A chapter entitled “Language Evolution with Deep Learning” will soon appear in the yet-to-be-published “Oxford Handbook of Approaches to Language Evolution.” It describes experimental approaches to the emergence of language in humans using AI. The authors apply gaming principles to show how machine learning and neural networks can be used to develop language through trial and error. Gaming presents a problem that requires an AI to solve a task. A combination of supervised and reinforcement learning lets the AI begin to understand the rules of playing the game.
The authors of the chapter present the Visual Discrimination Game involving a sender and a receiver. The sender sees an image and communicates it to the receiver. The receiver based on the message guesses what the sender is seeing before the the image is revealed to it. Played repeatedly, this communication over time eliminates errors and retains successes producing an image and communication dataset that becomes increasingly sophisticated and discriminating over time. Although not a conversation, it represents an avenue of communication. The Game doesn’t replicate human language but it represents a means to more complex learning involving other aspects of language including conversations and interactions with the physical world.
It is this type of deep learning advancements that will lead to AI discovering the fastest route to create realistic language evolution scenarios at unprecedented scales and to the building of intelligent language models. It will give us the means through AI to decode messages from other species here on Earth and eventually communicate with them, and it should allow us to do the same with extraterrestrials. Employing AI deep machine learning and neural networks may make it possible for us to do what Dr. Doolittle could do, “talk to the animals.” As for other aliens here on Earth, and I’m speaking about AIs themselves, as well as those who reach us from the stars, AI may unlock the key to intelligent conversations with intelligence in the Universe that find us or we find in the future.