Re: Cowan on morphology

[Mr Andrew Rosta <ucleaar@UCL.AC.UK>]

> I don't see how our morphology for compounds has anything to do with
> learnability of >gismu<.  You can learn gismu without learning either
> rafsi or lujvo.

By "gismu" I meant "the grammar of gismu", i.e. meaning, phonological
structure, morphology. There is a distinction between morphological
gismu & "lexical" gismu. I don't know what the proper term is for
the category that includes a gismu & its associated rafsi.

> I can't speak for Nick or Colin, but I suspect that most often, people
> arew writing without particular regard for their audience, or perhaps
> to write at an audience that they know will be using the word lists,
> or to write for their peers in the use of the language, who have a
> reasonable chance of knowing the rafsi in question.  I agree that it
> would be nice to have more texts aimed at beginners, but people don't
> write such texts.

I agree. I meant simply that although it is *possible* for peopleto
use 5-letter rafsi, they don't in practise. I'm not suggesting
people *ought* to use 5-letter rafsi.

> My statement about use of rafsi applied to the experiences we have had
> here in LIVE conversation, both conversation sessions, and at LogFests.
> Very few lujvo, and the ones that are used are composed of rafsi that
> are well known (like 'sel-' for se conversion, etc.), and I myself use
> some expanded lujvo, when creating them on the fly - OR, if I get a blank
> look, I expand it immediately for the listener.  I think the conversations
> on the IRC have also minimized lujvo.

But is the use of expanded rafsi due to speakers being considerate, or
to speakers not being fluent enough to *use* shorter rafsi?

> A lujvo based on klama in final position should have something to do with
> 'klama'ing. Now the language won't always be under prescriptive control,
> but while it is, I suspect that no 'sapphire' ending in 'klama' words will
> get into the dictionary.  Indeed, at the moment within the community, there
> remains a very strong literalism trend that objected to the relatively
> lesser sloppiness of JCB, who used 'zmadu' (x makes why from z) for causals
> in a very malglico manner.  The standard that we teach is that a lujvo
> should represent one specific meaning from among the possible meanings that
> the associated tanru would have, recognizing that some amount of tanru
> modification could take place to bring places fromt he modifier terms into
> the lujvo.  The primary debate has actually been whether the determination of
> such place structures should be more or less algorithmic from the source
> tanru, which practice has NOT been accepted.  But Nick's writings on lujvo
> making are promoting a standard only one step less drastic.

I'm just going by what the grammar says. If the above paragraph represents
official policy, the rules on lujvo formation should be tightened.

> Loglan pre-GMR was very much like what you suggest would be better -
> allowing jbama and klama to both be represented by -ama in a compound.
> It was not as you say - people had to memorize every word they wanted to use,
> and to rely on the dictionary for every little thing they did.  The result
> was that there was far less Loglan text written than you see these days
> being posted to Lojban List.  And people DIDN'T like it, and they complained.
> And one noted linguist (Zwicky) was especially critical of this.

I think this would depend on the compounding rules, & also on how
phonologically distinct gismu are from each other. I didn't actuallysuggest
that allowing -ama to represent _jbama_ & _klama_ would be better;
I suggested that regular compounding rules would be better, & cited
-ama as an example.

In Zwicky's 1969 _Language_ review I cannot find a criticism of Loglan's
morphotactics. I can, however, find a criticism (p448, top) of the
lack of semantically regular, grammatically prescribed word
formation rules.

----
And