Message-Id: <m0tDFYZ-0000ZTC@xiron.pc.helsinki.fi>
Date:         Wed, 8 Nov 1995 11:41:37 -0500
Reply-To:     Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Sender:       Lojban list <LOJBAN@CUVMB.BITNET>
From:         Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Subject:      response to Dylan on rafsi
To:           Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 2748
Lines: 56

Dylan answer mark (quite well, BTW):
>> The most important of these complicating factors is the process of
>> learning Lojban.  Homophone affix ambiguity makes that process more
>> difficult, primarily because of one glaring inconsistency:  some rafsi
>> have the same meanings as their identical twin cmavo, whereas other rafsi
>> do not.
>
>Well, I suppose the official answer is that you're not supposed to learn
>rafsi until fairly late in the game, instead learning lujvo as units.

Certainly not by policy or pedagogy.  You certainly learn many cmavo
before you learn your first rafsi.  But I suspect that I would learn all
the rafsi before I mastered all the cmavo.  Some cmavo are pretty
obscure.

Note also that all cmavo that have rafsi are taught as part of the gismu
list in LogFlash, so in fact you have almost certainly learned all THOSE
cmavo before you have learned any rafsi.  Of course, you also have been
exposed to the rafsi for each of those cmavo, and thus could presumably
notice which of them actually have their own image as rafsi (but I
certainly have never taken notice of this).  I just looked and there are
25 CVVs that have themselves as rafsi, around 5% of CVV rafsi.  I'm not
sure I have seen any of them used in lujvo yet, and the few I think are
most likely to be so used are unlikely to suggest many other meanings
(for example, 21 of the 25 have the letter o in them, and the letter o
is found in only 225 gismu at all, half of which are
cultural/metric/element words).  Thus it can almost be ASSUMED that a
rafsi does NOT mean its associated cmavo, and memorize the exceptions
if/when they occur.


>...  This
>> guarantees that any lujvo containing one or more of the 295 ambiguous
>> rafsi will be resistant to comprehension by the learner.
>
>But only when the ambiguous rafsi occurs at the beginning of the word,
>right?  I think these sound ugly anyway--I don't like the glue.  People
>seem to avoid them.

And not even then, because an r glue consonant cluster comes immediately
after the CVV (except in the rare CVVCCV lujvo, where the stress pattern
will make it clear that the CVV is a rafsi).  Since there are NO valid r
initial CC permissible initials, you have things resolved before the
last syllable of the cmavo ends (the 'r' usually gets glommed onto that
syllable in actualy speech, rather than left as an unstressed syllabic
'r' standing in its own syllable, which is the alternative).

>Oh, I just thought of one criticism of your proposal. rafsi of the form
>CVV or CV'V are, I believe, most aesthetically pleasing at the end of
>words; but cmavo in lujvo compounds rarely (never?) occur at the end.

Not never, but certainly rare. moi/roi/mei will of course be usually at
the end.


lojbab