Print

Print


Isn't this what the semantic web is all about? Or Google's Knowledge Graph?

stevo

On Sat, Jun 16, 2012 at 1:42 PM, Yaesen Kole <[log in to unmask]>wrote:

> Hello everyone !
>
> I want to talk to you about my project: A while ago, I wondered (for a
> story) what form would a language take if it was designed to communicate
> with aliens. One characteristic immediately jumped in my mind: a vocabulary
> as small and exhaustive as possible.
>
> So I wondered how many root concepts could be isolated in a list (in a
> conlang, all the other possible words with a link with those words could be
> obtained by adding affixes to the root word). I couldn't find such a list,
> so I decided to make one.
>
> But (and here is the technical part) to make such a list, I had to use an
> object-oriented approach. Think about it:
>
> If we take the example of an object (class "artificial object"):
> An artificial object has:
>    _a user (class "person")
>    _a maker (class "person")
>    _a use (class "action")
>    _a verb for using
>    _a verb for making
>    _ etc...
>
> Eventually, the idea is to take a large list of words, sort them in a
> "class tree", and after that, link them with each other to have kind of a
> "concept network". That will help me isolate basic concepts.
>
> To do this work, I'm using a list of words (a HUGE list, ~150k words (with
> nouns, verbs, adjectives, etc...), necessary for exhaustiveness), all in
> French (my native language), that I reduced from ~30k words to ~15k words
> for nouns (didn't process verbs or adjectives or anything else for now) by
> removing redundancies and invented words you'd never see in a conversation
> (the list was made out of lots of books, with statistics that I will use
> later). It's VERY long, but I really want to finish it.
>
> This is a work that has been occupying my free time for about 2 years now,
> and I was wondering if anyone here ever attempted a similar approach? if
> you did, could you show me your list of classes? could you tell me about
> your results? I intend to publish mine of course, when I'm ready... but I
> have some tricky little problems to solve before that!
>
> Friendly,
> Yaesen
>