Naming Game

After my rant about linguistic coherence (see Savvy?), I did a bite-size portion of research about how new words become accepted.

I found the following question: What processes can explain how very large populations are able to converge on the use of a particular word or grammatical construction without global coordination?

Now, before I continue, know that this is excerpted and edited from a research paper (credits supplied below) so unless you are interested in words, you may find this a tad tedious.

(Hence 'cool' becomes 'book' per the fabulous Stephen Fry Qi)

The Naming Game is played by a population of agents trying to bootstrap a common vocabulary for a certain number of individual objects present in their environment, so that one agent can draw the attention of another one to an object, e.g. to obtain it or converse further about it.

Each player is characterized by his inventory, two players are picked at random and one of them plays as speaker and the other as hearer.

Their interaction obeys the following rules:
Speaker selects object -- Speaker retrieves word from inventory, or, if inventory empty, invents new word -- Speaker transmits selected word to hearer -- If hearer has the word in inventory, the interaction is successful, both players keep only the winning word, deleting all others -- If hearer does not have the word, the interaction fails and the hearer updates inventory by adding an association between new word and object.


Then there are a few assumptions, some really nifty equations, S-shaped curves and a passing celebratory reference to Zipf's law. (You've made it this far, don't give up on me now!)

Ain't linguistics book?


Credits: Baronchelli et al.
Sharp Transition towards Shared Vocabularies in Multi-Agent Systems