From pbarreto@unisys.com.br Tue Jan 16 20:03:38 1996 Received: from panther.unisys.com.br (panther.unisys.com.br [200.255.218.10]) by locke.ccil.org (8.6.9/8.6.10) with ESMTP id TAA28139 for ; Tue, 16 Jan 1996 19:20:48 -0500 Received: from atsp12.unisys.com.br (atsp12.unisys.com.br [200.255.218.20]) by panther.unisys.com.br (8.6.12/8.6.9) with SMTP id VAA11043; Tue, 16 Jan 1996 21:58:10 -0200 Message-Id: <199601162358.VAA11043@panther.unisys.com.br> X-Sender: pbarreto@pop.unisys.com.br X-Mailer: Windows Eudora Light Version 1.5.2 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Tue, 16 Jan 1996 21:54:08 -0300 To: John Cowan From: Paulo Barreto Subject: Re: A little theorem on a forbiden fu'ivla form Status: OR X-Mozilla-Status: 0011 Content-Length: 1151 >> So far I wasn't able to disprove the validity of {spra'i}, or more generally >> CC...CV'V fu'ivla. > >The specific form CCCV'V can't be a slinku'i failure, for there is no >way to segment CVCCCV'V into rafsi: CVC+CCV+'V is not possible, nor is >CVCC+CV'V. > >As for the general form CC...CV'V, "slinku'i" itself is of this form, and >"pa slinku'i" segments as "pas+lin+ku'i". So the form is not safe in general. Oops, I failed to express the suitable regular expression. I mean a sequence of three or more consonants, followed by V'V. That is, CCCC*V'V. This form seems safe. >> I have a further question on fu'ivla morphology. Is it possible that a fu'ivla >> be stressed on vocalic consonant (as in XRvatska, for instance)? > >No. All brivla have penultimate stress, and syllabic consonants (and "y") >do not count. And I just noticed that even {xrvatska} is not valid, as the initial consonant cluster contains a pair (rv) that is forbidden at this position. Paulo S.L.M. Barreto -- Software Analyst *** PGP public key available on known keyservers *** e'osai ko sarji la lojban