To: lojban-list
Subject: responses to various technical issues raised
Date: 29 May 90 03:02:28 EDT (Tue)
From: wetblu!uunet!cbmvax!snark.uu.net!lojbab
Message-Id: <9005290302.AA11452@snark.uu.net>
Status: RO

Subject: misc technical responses to questions
to a. protin, re the response from jim carter
  Jim's response was more or less the same as mine, though I don't
  especially like the translations he gives.  However, his example from
gua!spi points up a distinction that I did not make.  Neither
   da barda nanmu      nor da nanmu barda
means what Jim's version says, which is "something is big and a adult-male-human"
The latter would be expressed in Lojban through use of a logical connective:
   da barda je nanmu
The essence of Lojban tanru is NOT logical, but is a semantically ambiguous
'metaphorical usage' (pardon me John C. - I'll get to this in a moment).
Theoretically, any tanru should be re-expressible in some number of additional
ways to make it clear just what the 'modification' the first term is making
to the second term.  Thus my interpretation of da barda nanmu could be
more explicitly expressed using
   da nanmu gi'e barda leka nanmu
"something is a man, AND is big in dimension/property adult-male-human-ness"
whatever such property might be defined as (you could specify other places
such as height and weight, and use man-ness as the 'standard' (x3 place) that
'da' is big for"  but this is getting complicated.

turning to J. Cowan's writings, of which there have been several good ones
   1. on keyword proposals for tanru and lujvo - I'll present your proposals
LogFest for a fair hearing because your points of definition are valid.
However, I personally am opposed to the changes, based on my intent for
English keywords.  These are NOT inended to be literal translations of the
Lojban, but are supposed to me short memory hooks to hang the Lojban word on.
Key is 'short' and your proposals are NOT 'short' keywords.  Secondly, they
are jargon that will mean absolutely nothing to any non-linguist, i.e. the
majority of the community - they give no clues as to meaning.  Even if you know
to use the linguistic concept of 'compound', most people will not accept a
multi-word non-hyphenated construct as a 'compound', whether you qualify it
as 'open'.  In short, the linguistic jargon is itself metaphorical.  The book
I'm reading on lexicography use different terminology, by the way.  I suspect
that there really isn;t a good English equivalent for tanru - which is why we
use the Lojban in talking about the language.
'metaphor' for all its faults and ambiguity, uniquely describes the major
distinguishing feature of tanru, an ambiguous, metaphorical relationship
between two concepts, that is merged into a gestalt.  your definition that
'metaphor' involves the transfer of meaning from one concept to another, which
is what happens in tanru.
The fact that a metaphor is literal rather than figurative does not make it
any less a metaphor.  Just because many American poets don't like metaphors
that are analytic does not mean that such things are not metaphors.
   A minor point against, by the way, is that keywords are also meant to be
short for the purposes of typing them in for LogFlash.  Especially when you
want to use a non-obvious keyword, making it long, means only that more
people will miss it through typos.  It will not make them better understand
the concept.  A better bet would be to clarify the long-form definition of the
word tanru (and lujvo) to include the technical linguistic jargon.
   We are trying to move AWAY from the use of linguistic jargon in our teaching
materials, with the original stimulus being Nancy Thalblum, who you are
studying with.  Ask her opinion on the jargon keywords, and be prepared for
strong disagreement.
   Nora, by the way, like the term 'modification pair' as a fairly clear non-
jargon explanation of tanru.  I would favor trying this in the longer English
definition, and trying it in teaching materials that I and others write,
leading to adoption of a baseline change only if a real benefit obtains.
   I am in general opposed to the use of the word 'compound' in any of our
keywords, by the way.  It has turned out that this word, like 'predicate'
just crops up too many places with too many disparate meanings.
   Deviating slightly, a proposal has been made to change the keyword for
'gismu', from 'primitive' to 'root word'.  I'm going to support this one,
which counters a misconception and is clear to non-linguists (It's already
been run by Nancy T.)

  2.  On 8 interpretations of an English sentence.  There are actually several
more, and you made some errors.
  a.  use gi'a instead of bu'a for logically joing within sentences
  b. the veridical/non-veridical distinction is significant, but is not a claim
of existence.  You can say the equivalent of 'lo unicorn', without making a
logic error.  To claim veridical usage AND to claim existence, you use the
form "da poi finpe" for "something that is a fish".  To say you are seeking
"lo finpe", says that you are seeking "something that really is a fish",
while you are correct in that "le finpe" need not even be a fish at all (it
could be a dolphin).
   Thus, you have three possible expressions for either of the two terms
'fish' and 'bicycle', or for both.  In addition, you have the logical
connective dichotomy between exlusive and inclusive OR.
  c. The logical connection of two clauses with sisku (the one using *bu'a for
gi'e) is identical in meaning to the logical connection of the sumti alone.
I'm not sure what the difference that is trying to be conveyed is (there was
mention of John vs. the observer), but this doesn;t achieve it.
  d. I vaguely suspect that there are a few other possibilities for the English
besides those of the nature you've mentioned.  The mention of 'intensional'
and 'extensional' (more linguistic jargon word that seems to have several
meanings depending on context), suggests to me that you could play around with
tenses.  'seeks' is a present tense, but in Lojban, this is distinguished
not been transferred to here yet (and I'm not sure I understand the comment)
from the tenseless version. (your translations did not make this distinction)
But even in English 'seeks' could be taken as continuous, intermittent,
habitual, etc. activities.  Since the original question was 'natural language'
you would also have to explore various alternative tense systems not found
in English.
  e. Noting also some discussion on sci.lang by someone named greg lee, which
ought to be relayed to this list (though I'm not sure I understand the comment).
Lojban allows indefinite numbers of levels of quantification scopes, so I
believe we can handle any such problems. I'll run the comments by pc at LogFest.

   3. On the letteral proposal.  Much of it looks good, though some of your
unused cmavo are already taken.  I like the idea of dual-letter letterals,
though I think it needs to be tested by using it with a couple of complex
alphabets.
  Responses to some of your questions/comments, though.  The list of alphabets
is not arbitrary, but is selected on the same basis as the gismu list cultures.
They are the major alphabets used by our 6 source language cultures.  Greek and
Hebrew were included because they are used in mathematics and physics (as well
as religious writings).
  The Chinese do not use an alphabet.  Their symbols do have a classification
scheme that can be patternable on the letteral system, but I don't have the
know-how to do it, so I left it out with the intent of having someone who knows
Chinese help out.  Japanese was left out because it wasn't one of the basic 6.
We recently added the International Phonetic Alphabet to the list, by the way.
  JCB originally had only Greek and Roman alphabets represented, and each uses
a cmavo in 'prime cmavo space'.  The reason for including non-Lojban alphabets
is NOT for use in spelling Lojban (which seems a waste of time with totally
phonetic spelling).  Rather it was for use in MEX and scientific usages.  I
broadened this to include spelling words from other languages that are not
phonetic, but this is a secondary goal.
  In the existing set, there is provision for an 'auxiliary' set.  This is
for punctuation marks and the like, which should be easily distinguished from
alphabetical characters.  At the time of the last cmavo list update that
dealt with letterals, we still though of the period and comma as punctuation.
I agree that they should be assigned some type of alphabetic value now, though.
  The miscellany of grammarless lexemes exists because until MEX is complete
(and letterals are considered primarily as part of MEX), we won't be sure that
there isn;t going to be needed some grammar for one or more of these various
shift operations.  Thus shifts and cancels will remain in multiple lexemes
for a while longer.
  We'll have to look closer at other aspects of the proposal in light of our
differing purposes.  I suspect that much of the proposal will be adopted, but
not all.