Date: Wed, 28 Dec 1994 23:34:04 -0500
From: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Subject: terse tense
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
Reply-to: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Message-id: <01HL7BGFRAXY000VY6@FIPORT.FUNET.FI>
Content-transfer-encoding: 7BIT
To:         lojban@cuvmb.cc.columbia.edu
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Content-Length: 4895
Lines: 102

Gerald Koenig writes:
>Subject:      quantified terse tense

and gives lots of examples towards his attempt to devise a short form
for specifying the length of a fixed interval in time.

There are several problems with his ideas.

1. The use of "vei" in his examples adds nothing to the sentences in question.

Thus in:
>New usage:
>le nanmu zu'a vei ci mitre cu batci le gerku
>The man [left] [three meters] bites the dog.
>
>({<le nanmu KU> <zu'a [(vei {ci BOI} VE'O) mitre KU]>} cu {batci <[le
>gerku KU] VAU>})

leaving out the "vei" does not materially change the parse:

({<le nanmu KU> <zu'a [{ci BOI} mitre KU]>} cu {batci <[le gerku KU] VAU>})

I won't try to remember-and-explain the purpose of vei off-the-cuff, but
whatever it is, I don't see what effect Gerry is trying to achieve by
it.  This may be complicated by other problems.

2. As Cowan has pointed out "ci mitre" does not mean "3 meters" as
commonly understood, but rather "3 meter-length objects".  It thus does
not convey any sense of interval.  For that you need "mitre cimei", or
"mitre be li ci".

3. The goal he is trying for seems quite fuzzy to me.  I am gathering
that he wants to find a short expression for something he believes is
expressed briefly and unambiguously in English.  But the English is
seldom all that unambiguous.

For example "3 years ago" does not generally mean that one should move
exactly 3 years in the past from now.  It may mean a time interval
encompassing that point in time, or a time interval immediately
preceding that point in time but not including it.  English tenses are
usually by implicature non-aorist "3 years ago I was employed" implies
that I am not currently employed.

Likewise "stand 3 meters to the left of me" may mean that the midpoint
of your feet should be 3 meters from the midpoint of mine, or that the
shortest distance between our feet is either exactly or at least 3
meters, or it may mean that no part of our bodies should be less than 3
meters apart.  Spatial tenses in English are definitely non-aorist.
"the road runs 5 miles from here" forbids that the road be any closer
than 5 miles, allowing for sokme fuzziness as to what 'five miles'
means.

4.
>Furthermore the resulting expressions need to be concise enough to
>equal natural language.

This appears to be a restatement of something Gerry wrote earlier:
>I believe that for lojban to be more than a toy, it must equal the best
>languages of the world in conciseness as well as exceeding them in
>precision and consistency.  It is hard for me to believe that there are
>unbeatable tradeoffs here.  Optimization can produce a language with the
>best of both worlds.  We are at the beginning.  Either we do it or
>another language will be spawned.

But Lojban is not designed for optimized conciseness - it isn't anywhere
in our constellation of goals.  You are asking Lojban to be BETTER than
every language in two design features, and equal to the best in a third
(assuming that "best" is something that can be agreed upon, which is
doubtful).  But you have left out other goals that are more important to
Lojban:  metaphysical parsimony, logical structure.  And you are
fighting Cowan's oft-repeated maxim about the price of infinite
precision being infinite verbosity.  The reduced redundancy of Lojban's
tightly-packed word-space isn't even enough to counteract the density
natural languages achieve with ambiguous and multiple meanings for words.

So, without going into this further, I don't agree with the goal.
Lojban cannot be as concise as natural languages in many of its aspects
- for example, its logical connectives and prenex structures are not as
concise as natlangs, in part because natlangs do not have the need to
express some distinctions that Lojban permits or even requires.

But even if I did accept the goal:  which natural language?  Not all
natural languages are equally concise in expressing fairly complex
ideas.  English seems very concise in these expressions, but other
natlangs are not as concise.  Those other languages may be more concise
in some areas where English is wordy.

pc has said that most tense structures are really an abbreviated form of
subordinate clause, which means that any given tense structure can be
expand into a second sentence.  There is a tradeoff in tense grammars
between conciseness and expressivity - there will always be *some* tense
concepts that will have no short "tense" expressions form.  But Lojban's
tense structure probably exceeds that of any natlang in total
expressivity - the variety of tense concepts that can be expressed
briefly.  The fact that this means that some structures aren't as brief
as their more ambiguous English counterparts does not bother us greatly.
It is safe to say that Lojban provides more variety of concise
expressions than any single natlang (even if not all of them are
especially useful).

lojbab