Scientists have tried contacting extraterrestrials with a number of bespoke linguistic systems. But we might be better off using our own languages.
In May 2018, a radar facility in Tromsø, Norway, trained its antennas on GJ237b, a potentially habitable exoplanet located 12 light years from Earth. Over the course of three days, the radar broadcast a message toward the planet in the hopes that there might be something, or someone, there to receive it. Each message consisted of a selection of short songs and a primer on how to interpret the contents.
This was the second iteration of Sónar Calling GJ273b, an interstellar messaging project by the nonprofit METI International that began in 2017. Although both transmissions were billed as a “music lesson for aliens,” the second broadcast was notable for rehabilitating an extraterrestrial language developed by the physicists Yvan Dutil and Stephane Dumas in the late 1990s.
This custom symbolic system begins by introducing ET to numerals, and then progresses to more complex topics like human biology and the planets in our solar system. An earlier version of the language was first sent into space in 1999 and again in 2003 as part of the Cosmic Call messages—a crowd-sourced interstellar messaging project that marked the first serious attempt at interstellar communication since Carl Sagan and Frank Drake sent the Arecibo message into space 25 years earlier.
All of these formal messaging attempts have taken basically the same approach: Teach numerals and basic arithmetic first. But as some recent insights in neurolinguistics suggest, it might not be the best way to greet our alien neighbors.
The world’s first interstellar communication system, the lingua cosmica, or Lincos, set the tone for all subsequent attempts by placing basic math at its core. Designed by the Dutch mathematician Hans Freudenthal in 1960, Lincos inspired several other mathematicians and scientists to try their hand at designing extraterrestrial languages. Each system is ultimately an attempt at solving a remarkably complex problem: How do you communicate with an intelligent entity you know nothing about?
The question gets at the nature of intelligence itself. Humans are the only species on Earth endowed with advanced mathematical ability and a fully fledged faculty of language, but are these hallmarks of intelligence or human idiosyncrasies? Is there an aspect of intelligence that is truly universal?
Scientists and mathematicians have grappled with these questions for centuries. As the Nobel laureate Eugene Wigner once observed, mathematics is “unreasonably effective” at describing the natural universe, which has led a significant contingent of mathematicians to conclude that math is baked into the fabric of reality. From this perspective, mathematics isn’t something produced by the human mind so much as something the human mind discovers.
Most interstellar communication systems were designed around this conclusion. The goal isn’t to teach ETs about addition and subtraction—presumably they know as much if they can build a telescope to receive the message. Instead, these systems teach ETs about the way we code numbers as symbols. Then they can build up to more complex ideas.
It’s an elegant solution to a difficult problem, but Lincos still rests on the assumption that an ET is “human-like in its mental state,” as Freudenthal once conceded. But if ET does in fact think like a human, does that alien also have some kind of human-like language?
That was where Marvin Minsky and John McCarthy, two of the progenitors of artificial intelligence, landed after they became interested in interstellar communication. Both Minsky and McCarthy had a deep interest in the search for extraterrestrial intelligence, which they realized had a lot in common with their own search for artificial intelligence. As Minsky argued on several occasions, ET is likely to have language because language is an ideal solution to the fundamental problems faced by any intelligent species—namely constraints on time, energy, and resources.
A deeper question is whether ET’s language will be similar to our own. In other words, whether it will also obey the universal grammar, the hierarchical, recursive structure that linguist Noam Chomsky has argued is the deep structure common to all human languages. Although languages tend to be analogized as a form of software running on the hardware of our brain, recent work in neurolinguistics suggests that language—and the universal grammar—is actually an expression of the hardware itself.
Several brain imaging studies have shown that the deep structure of human language manifests in our neural activity. When people are taught fake rules for either a made-up or real language, their brains respond differently than when they use actual languages (whether familiar to them or not). These findings suggest that the shared attributes of natural languages might be encoded in how neurons connect. In other words, our faculty of language may be inextricably tied to the structure of our noggins.
If extraterrestrials do have a language similar to ours, that might imply they also have a functionally equivalent neurobiology. To say aliens might think like us and have language is one thing, but to argue they have brains like ours pushes the limits of credulity. But it might not be as crazy as it sounds.
Biology, after all, is beholden to the laws of physics, which puts constraints on the trajectory of evolution. Astrobiologist Charles Cockell makes this argument in his recent book, The Equations of Life, in which he points to the remarkable similarities across species on Earth—from the fact that life is cellular and arises from the same four nucleotides, to the structure of an eye or a wing. This is not to say that evolution is deterministic (random events like asteroid impacts and genetic mutations still happen), but that the number of evolutionary end points isn’t limitless. In other words, we’re not going to discover a planet inhabited by sentient ice cubes.
There is a good chance that ET’s planet will be quite a bit different from our own, and the species there will adapt accordingly. But the course of evolution on ET’s planet will still be bound by the same physical laws, and ET will face the same fundamental constraints on time, energy, and resources. So it is reasonable to assume that extraterrestrial evolution might arrive at similar solutions to these common problems, such as a brain capable of wielding hierarchical, recursive languages.
If that’s the case, then the best way to communicate large amounts of information may not be painstakingly designing artificial languages from scratch, but sending a large corpus of natural language text, such as an encyclopedia. This is how we train natural language algorithms on Earth, which tease out the rules of human language by statistically analyzing large collections of text. If ET has developed its own AI, it could potentially decipher the structure of a natural language message.
Of course, natural language processing algorithms on Earth don’t really understand the meaning of the text they analyze. They are blindly manipulating symbols. And aliens may still need some kind of extraterrestrial language to connect some of the symbols of human language to their meaning. But as on Earth, the best way to start an interstellar conversation might simply be saying “hello.”