Hacker News Reader

While I agree we've no way to know if Zipf's law applies to alien firms of communication, it seems extremely unlikely that it would apply to random or natural signal profiles that were not communicating information.

So failing to conform to Zipf's law doesn't prove a signal is not meaningful, but confirming to it is likely to be very strong evidence that a signal is a form of communication.

View on HN

> So failing to conform to Zipf's law doesn't prove a signal is not meaningful, but [conforming] to it is likely to be very strong evidence that a signal is a form of communication.

Then again, Zipf's law is known to describe many non-communicative, "natural" distributions like the sizes of grains of sand, the sizes of asteroids, and the populations of cities (one of Herbert Zipf's original objects of study). And conversely, many levels of representation of human communication have non-Zipfian distributions, like phoneme distributions. (See my comment history for references.)

Reply on news.ycombinator.com