Once I saw an experiment on a documentary which a simple artificial language (just a set of ten words or something like that) was artificially created by a recurring process of memorizing errors of the previous person on remembering the given language (which was initially just randomly created) and the result of each iteration was passed to next person, which after some iterations, the result was a language that was easy to memorize and also some incidental rules could be extracted from it. What method is called that?

1 Answer 1


This sounds like one of the series of papers by Kirby and/or Smith; e.g., Smith, Kirby, Brighton 2003. They just call it 'iterated learning'.

  • 2
    That's a model. I think Kirby, Cornish, and Smith (2008), which used actual chains of humans, is a better reference. pnas.org/content/105/31/10681.abstract Jun 15, 2016 at 4:16
  • @RussellRichie Thanks. I just grabbed one to show the term, but that is indeed the paper. Jun 15, 2016 at 4:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.