Re: Bootstrapping a cooperative conlang

--- Herman Miller <hmiller@...> wrote:

<snip>
>
> So is the intended purpose of this system a way of documenting the
> various usages of a word in a particular variety of language? Or are
> there various parts of the definition that explain how the meaning of
> the word varies over time and in different dialects?
>
If by "system" you mean NSM, then I suppose that is the purpose.

If by "system" you mean the collaborative conlang system I'm trying to build
then most emphatically NO. What I have in mind is "growing" a new conlang by a
process that complete novices could participate in. After all natlangs were not
invented by linguists, but by common folk talking about common things. I'd like
to grow a conlang that way by letting linguistically naive people collaborate
as well.

<snip>

> If you've identified a group of potential collaborators who are all
> familiar with Wierzbicka's work, and who have the time and level of
> interest to write definitions in that style for all the words you're
> going to need, that could be an interesting thing to try. I just think
> that having to learn a technical set of words for basic meanings, and
> how to use them, is likely to be intimidating to just about anyone who
> might want to participate.
Heaven forbid that contributors would have to be familiar with something that
esoteric!

> Say that you want to define a word "fur" to use in your definition of
> "mouse". Looking at the semantic primitives, I don't really have much of
> a clue where to start. Probably "fur" is too high a level and it needs
> to be broken down into something like "dense hair", where "dense" could
> ultimately be built from "near" and "touching" and "many/much" in some
> way. There's still a big gap between the semantic primitives and
> something as fundamental for animal definitions as "hair".
>
"hair" and "fur" would probably end up being illustrated with a picture, or
roughly described as "warm, fluffy animal covering", or "stuff that grows on
top of people's heads", accompanied by a picture.

<snip>
> An illustrated index of defined words would be useful, but if
> you've got that, the illustrations alone might be a better place to
> start with.
Absolutely! Pictures will play a big role in the process for every word that
can possibly be defined by a picture. My preliminary dictionary database
http://fiziwig.com/tazhi/view_def.php already has the ability to assign images
to words.

> Say that someone wants a word for "guitar" -- you first need definitions
> for words like "string", "fret", and various other parts of the
> instrument, along with a description of the general shape and size. Or
> you could just take a picture of a guitar and upload it. Then if there's
> enough interest that it becomes important whether a ukulele or a banjo
> falls under the definition of "guitar", there could be a discussion
> which ends up producing a more detailed definition.
First of all, a picture of a guitar would all that is needed. Secondly, the
last thing I want is an in-depth discussion of how to define "guitar". After
all, the purpose of the project is to have all discussions IN the new conlang,
not ABOUT the new conlang. All I want is for someone learning the language to
know what "guitar" means. How they get to that knowing should be simple,
straightforward and practical, and understandable at a glance by the
linguistically naive.

In other words, I don't need to tell anyone what a guitar IS, I assume they
already know that. I just need them to learn the WORD for the THING they are
already familiar with.

--gary