On Thu, 3 Jun 2010 14:20:15 -0700, Lee <[log in to unmask]> wrote:

>In the Artificial Languages section, the article contains a link to an
"alien language experiment":
>(I played the game, but I don't understand the point. Are there supposed to
be pictures of alien fruit to go with the sequences?)

No, that's a different experiment, just a Simon type game with the same
principle.  Unless there is no tree (and the tree did get cleared once while
I was playing with it), its prompts are never completely new sequences, only
ones stored in the tree that someone has previously given it as a response.
 I assume it's the same tree across all users.  

It was illuminating to play it with the tree and game windows open
simultaneously:  after submitting a sequence in the game, a new node will
soon pop into existence in the tree, colourised to reflect how correct you

Anyway, it will be interesting to see how they elaborate that to make it
reflect "evolution of culture".  

>It would be interesting to run the experiment on various conlangs, but
that's probably going way beyond the effort most people could realistically
put in. Maybe there's a way to encourage and accelerate a similar process as
it occurs by a conlang's creator? I wonder if that would require a conlang
to be sufficiently large enough that its creator loses track of existing
words and idioms, thus creating new ones for concepts already described.

Yeah, it's a very cool thing they managed to get in their first experiment,
the emergence of orthogonality (or something close) from completely random
initial conditions.  

It's very much like what Gary has tried to achieve with several of his
collaborative experiments.  I remember one of his recent suggestions was a
language induction game sorta like this, which he cooked up by hand an
example of.  And I had a variant which I'd still love to see implemented --
I forget whether I described it in detail then, but I gladly would now, if
there's interest.