Message-ID: <45622725.7010705@lojban.org>
Date: Mon, 20 Nov 2006 17:07:33 -0500
User-Agent: Mozilla Thunderbird 1.0.7 (Windows/20050923)
MIME-Version: 1.0
References: <492596e80611182014o7608f9d3k76f7247f5926b98b@mail.gmail.com> <1df90d2b0611182026yeb05a5eo95f815883cb20fd@mail.gmail.com> <836504149.20061119104734@mail.ru> <4560FB73.5060706@lojban.org> <4810154213.20061120071558@mail.ru>
In-Reply-To: <4810154213.20061120071558@mail.ru>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Errors-to: lojban-list-bounce@lojban.org
To: lojban@yahoogroups.com
From: Bob LeChevalier <lojban-out@lojban.org>
Reply-To: lojbab@lojban.org
Subject: [lojban] Re: ki'a
Content-Length: 2734
Lines: 63

Yanis Batura wrote:
> On 20.11.2006, 6:48, Bob LeChevalier wrote:
>>Yanis Batura wrote:
>>
>>>{pinpedi} is a stage 4 fu'ivla, like {djirafa}, {alga}, {caxmati} etc.
> 
> 
>>And is an example of why I am opposed to inventing stage 4 fu'ivla at 
>>this point in the language evolution - as we can see, no one has any 
>>idea what it means.  A fu'ivla should remain a Type III - marked by a 
>>semantic-hint rafsi - until its usage is so common that usage (and/or 
>>Zipf's law) indicates a need for a shorter form.  I don't think that any
>>fu'ivla has received such usage.
> 
> 
> There are only 1342 semantic hint classes.

Most natural languages using a semantic classifier system have a couple 
orders of magnitude fewer classes (they are commonly called "genders", 
Swahili has 14, and Russian has only 3), and they work anyway.

>So, if we want to fu'ivlize
> games like "Go" we will have to stick to {kelcr-}: {kelcrgo}. One
> might then wonder what type of PLAYER Go is.

1.  semantic classes are not tertanru.  There is no requirement that XXX 
as a semantic class means that the fu'ivla is a kind of XXX.  Rather it 
is an attempt to limit the semantic space one must consider in guessing 
the meaning of an unknown (and probably undefined) word borrowed from 
another language.  Of course if it happens to be a suitable tertanru for 
  a tanru expressing the concept, this can be a good thing in that the 
sentence might make some sense using the semantic classifier in place of 
the unknown fu'ivla as a rough approximation.  But I contend that making 
such an assumption is extremely risky since non-x1 places of the fu'ivla 
almost certainly would NOT match that of the classifier, even if the x1 
would.

2.  Since the place structure of kelcrgo would not be intuitively 
obvious in any event, maybe it should be OK if someone were to presume 
from a case tag "kelcr" that the place structure of kelcrgo is "x1 plays 
Go with x2 (with possible other places like board size and ruleset in 
addition).  There is of course no rule saying this has to be the case, 
but it counters the argument about what one might "wonder".

3. brivla express a predicate relationship which is not limited to the 
x1 place, and kelci as a semantic classifier can refer to any of the 
places of kelci as well as nu kelci, ka kelci, etc.

4. Incidentally, I suggest that one could also use jivna or zdile as 
semantic tags for the game of Go, depending on what they wish to 
communicate about the game.

lojbab


To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.