It is a popular belief that, given the immensity of our universe, we are unlikely to be the only intelligent matter. SETI and the frightening [METI](http://meti.org/mission) are premised on the idea that some of it might want to speak with us. The obvious next question is... using what language? A pragmatic xenolinguistics would estimate a probability distribution over all possible alien languages. When sending a broadcast we could aim to make it interpretable by as much probability mass as is reasonable. When intercepting a message we could condition on the attributes of the message to find a highest-likelihood starting point for our decoding efforts. I am not a pragmatic xenolinguist, nor a xenolinguist of any kind, so here’s an easier meditation: do languages have any necessary traits? Are there patterns entities must discover if they’re to have any chance of effective communication? --- A first guess: all communication channels are noisy, and noisy channels advantage digital communication schemes over analog schemes. Our word for "green" is not halfway in phonetic space between our words for "blue" and "yellow", but is instead an entirely unrelated word. Most animal communication schemes also appear to be digitized. For an absolutely incredible example read this: https://arxiv.org/abs/0912.4649 --- Some more reading: - "[fractal messaging](https://arxiv.org/abs/2403.06633)", sending messages over a timescale-agnostic carrier so they can be noticed by species who live much faster or much more slowly than we do - [covert communication](https://arxiv.org/abs/1506.00066), broadcasting an electromagnetic signal without detection, is in general difficult but becomes much easier when there is a lot of background noise (for example: your planet is in front of a star). The existence of these techniques implies any message we receive was likely meant for us - [xenolinguistics, towards a science of extraterrestrial language](https://www.routledge.com/Xenolinguistics-Towards-a-Science-of-Extraterrestrial-Language/Vakoch-Punske/p/book/9781032399591), I'm linking as an anti-endorsement, most of these chapters do not do a good job of engaging with the topic