From jorge@xxxxxxxxxx.xxx.xxx Fri Jan 8 07:32:48 1999
X-Digest-Num: 40
Message-ID: <44114.40.144.959273824@eGroups.com>
Date: Fri, 8 Jan 1999 12:32:48 -0300
From: "=?us-ascii?Q?Jorge_J._Llamb=EDas?=" <jorge@xxxxxxxxxx.xxx.xxx
Subject: Re: Croatian as a brivla


>>It is not valid because xrv- is not a permissible initial consonant
>>cluster.
>
>Checking the fu'ivla rules, it seems that you are correct - each pair in a
>cluster of more than 2 must be a permissible initial and rv is not such.
>
>> It would absorb any immediately preceding cmavo to form
>>a different fu'ivla.
>
>I do not believe that this is a stated criterion against a fu'ivla.

Stated or unstated it is true.

>The
>rule ias that if it is preceded by a cmavo it must not break up into a
>lujvo (or a lujvo plus). If a longer fu'ivla can be interpreted as a cmavo
>plus shorter fu'ivla, the latter would be the interpretation, as I
>understand the rules so that no absorption is possible.

What you say is also true, but it doesn't contradict what I said,
and does not apply in this case. Of course a valid fu'ivla will not
absorb a preceding cmavo! That would make fu'ivla quite useless.
But something that starts with a consonant cluster that is not a
permissible initial (thus not a valid fu'ivla) will absorb the preceding
cmavo. If you say {mi xabju la xrvatska}, as long as there's no pause
between la and xrvatska, it will parse as {mi xabju laxrvatska}.

> But lujvo-based do
>take precedence over fu'ivla based ones.

In what case? I don't think this precedence could ever take place.
Nothing can be interpreted as both cmavo + fu'ivla and
cmavo + lujvo at the same time, so there's no need to set a
precedence.

co'o mi'e xorxes