Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

[NGL] Translation: Ku Valak siya à Weksaforda

1 view
Skip to first unread message

Jack Durst

unread,
Oct 7, 1999, 3:00:00 AM10/7/99
to
---------- Forwarded message ----------
Date: Wed, 6 Oct 1999 17:10:27 -0700 (PDT)
From: Gerald Koenig <j...@netcom.com>
Subject: Re: Re: [NGL] Translation: Ku Valaksiya ą Weksaforda

From: Gerald Koenig <j...@netcom.com>

Hi Jack and Stephen,
This is in answer to a request for a clarification on vector by Jack.
SAD
>> _feeling_ of the *particular* speaker, without the need for anyone
>> ever sitting down and saying, "Okay, this is a fair interval to
>> roughly characterise as a "short" vector with regards to time in this
>> particular frame of reference, this is a fair interval to characterise
>> as a "medium" vector with regards to time in this particular frame of
>> reference..." and so on. {hekai'} is _completely_ subjective - all
>> that matters is what the _speaker_ _thinks_ about the interval. He or
>> she, by employing it, is communicating purely a feeling or an
>> *opinion* about a length (volume?) of time.
JD>
>Perhaps it's because I'm the least experienced with vectors, but I've
>never felt an objectivity of this sort in the length vectors. Perhaps
>Jerry could explain it better... I had always thought that imprecise time
>vectors were strictly from the referent's perspective as to the degree of
>inspecificity; just as the difference between a near and far past in my
>system is entirely based on the narrator's perception of events with only
>very general guidelines.

This is essentially my view. But I want to discuss the general
guidelines: I took from my previous experience with the loglan project
that the threefold division of space-time is a language universal. I
suppose their objective meaning varies from culture to culture and also
from situation to situation in a given culture. Further, I suppose that
the speaker's intent and experience change the meaning, a child might
have a different meaning than an adult. So there is a speaker's
contribution to the meaning and a cultural contribution. These VXT
vectors wouldn't likely be used with a number or coefficient in front
of them, which would quantify them and make them fully objective. But
they are not fully subjective either, by default.

If I say "I mailed the check a short time ago" there are some real
expectations set up. It is not the case that the short VXT time
vector, <leti>, is completely subjective like

>> xxxx {hekai'} is _completely_ subjective - all
>> that matters is what the _speaker_ _thinks_ about the interval. He or
>> she, by employing it, is communicating purely a feeling or an
>> *opinion* about a length (volume?) of time.

So, there is no claim as to the consensual reality of the length of the
interval. To convey the idea that a short time interval is "completely
subjective" one has to say in VXT that it is in-mind only for the
speaker. One way to do this is simply to declare a virtual interval:

[TVS and VXT forms are mixed in these examples]

wo am hulte bin lotig. [inbin].
A person was waiting a virtual long time.

Virtual timelines are not related at all to real time. They can be
compressed, expanded, have gaps, etc. A problem once a subjective time
has been introduced into a narrative is where and if it stops, its
scope. If it is not clear from context, <inbin> is used to get Dorothy
back to Kansas. I don't know how PVS handles this. It is important in my
view to mark shifts of "Worlds" in discourse.

The subjunctive declarative modal <ror> could be used:

Wo am hulte ror lotig.
A person was waiting longly irrealisly.
<Ror> normally scopes one or two sentences that follow it, so strictly
speaking the structure here hides an implicit proposition:
"A person was waiting, it was a long irrealis interval for the person."
"Wo am hulte ror loti."for above. <loti> "means x is_a long time interval"


<ror letig>, (imaginary short time interval), is equivalent to <inheka>.

TYPE 1.
ror::-irrealis x says suppose that "p" is imagined only, in-mind
only, and irrealis; ie "p" is nonactual or nonfactual.
Some contrary to fact conditions.
[ Lightspeed exceeded; 2+2=7]

If a less extreme claim is appropriate, one could use ROS.

<ros letig>.

First, some ordinary time examples with various participle functions:
mi am reyla <lIMoti>.
I was working longly.
mam reyla <lotisoem>
I was working longly.
mam reyla <lotig>
I was working longly.
reylaomob <lotig>
I was working, longly. (for a long time-interval.) Is this correct, Jack?

ROS SUBJUNCTIVE TIME, on the imaginary timeline:

mif tibe roSig letig.
I will come (supposedly realis/irrealis) shortly.
mif tibe roR(ig) letig.
I will come supposedly irrealis short-time-vector-ly.

mi am reyla <ros lotisoem>.
I was working for a supposed long time. OR
I was working, I was supposing that I was working for for a long
time-interval, I may or may not have been.

miam reyla <ror loti> nir ku in`hapcu.
I was working a supposed long time in the dream. OR
I was working, I was supposing that I was working for a long
time-interval in the dream, that was contrary to fact.
<ror> scopes "loti nir ku inhapcu"

<ros loti> allows for the possibility that although the time interval
is primarily internal and subjective, it might coincide with objective
conditions. This is more the "I don't give a damn if it is or isn't I'm
in charge here and this is how I feel about it" that I feel Stephen
expresses. (That is a completely ror quote, Stephen:) )

TYPE 3.
ros::-in-mind, x says suppose that "p" is imagined and possibly
maybe also factual. also real, ie, In-mind and possibly out-mind.
Type I subjunctive. [Patent applied for]

Further,these tenses are expressible by directly referring to the
fundamental set of time-atoms [*atik] without the use of vectors. By
the way I read today that the lastest estimate of the age of the
universe is 12 billion years.

Stephen seems to want to do this his way; it's his self-realization, I
suggest he just include <heka> with the main PVS proposal and it will
be up to the users to pick it up or drop it. Meanwhile the explorations
of tense are very valuable to all of us. I continue to believe VXT can
express any time relations in a concise and accurate manner. The modals
are quite powerful too; they're hard to learn but easy to use.

Jerry
;;;
JD>

--------------------------- ONElist Sponsor ----------------------------

Get EXPERT CONTENT at ONElist!
Join PROS&PUNDITS. For details go to:
<a href=" http://clickme.onelist.com/ad/prospun1 ">Click Here</a>

------------------------------------------------------------------------


Jack Durst

unread,
Oct 7, 1999, 3:00:00 AM10/7/99
to
---------- Forwarded message ----------
Date: Wed, 6 Oct 1999 20:15:03 -0700 (PDT)
From: Gerald Koenig <j...@netcom.com>
Subject: Re: Re: [NGL] Translation: Ku Valaksiya ą Weksaforda

From: Gerald Koenig <j...@netcom.com>

Oops. Please put <bim> for <bin> in my previous post.

bim::-sets up a virtual world of predicates. x declares that a set of
statements, a "modal set" are given a value of true for a world. It can
be an imaginary world, such as the world of the Pokemon.

bin means knows.

Jerry

--------------------------- ONElist Sponsor ----------------------------

Choose from a wide selection of high-quality newsletters at ONElist.
For details on ONElist's PROS&PUNDITS newsletters, click below.
<a href=" http://clickme.onelist.com/ad/prospun2 ">Click Here</a>

------------------------------------------------------------------------


Jack Durst

unread,
Oct 7, 1999, 3:00:00 AM10/7/99
to
---------- Forwarded message ----------
Date: Thu, 7 Oct 1999 11:38:01 -0700 (PDT)
From: Gerald Koenig <j...@netcom.com>
Subject: Re: Re: [NGL] Translation: Ku Valaksiya ą Weksaforda

From: Gerald Koenig <j...@netcom.com>

>
>From: Jack Durst <sp...@sierra.net>
>
>Jerry: Thank you for your enlightening explanation.
>
>One point, though, the division of spacetime is by no means universal; its
>non-division in many Native American languages was the very thing which
>prompted Benjamin Whorf to consider the Sapir-Whorf hypothesis.

Well NGL is very lucky to have you as a linguist major among its
founders. I bought a copy of The Languages of the World, I guess I'd
better start reading it. I'm pretty sure though that loglan did say that
the threefold division of space was universal. I guess they were wrong
about that. Thanks much for the mini-lesson.

>
>In Navaho the past-future division is non-existant on the grammatical
>level, they effectively speak in what would be NGL generic tenses. In Old
>English the distinction was twofold past/non-past with no aspect marking.
>In biblical Hebrew the distinction is purely on aspect.
>
>Likewise, time and space are not always seperated, some South American
>languages use something like an underdeveloped VXT system wherin space and
>time are treated by the same inflections. One might mark
>past/present/future, but just as easily mark up/down/back/forward/
>or left/right instead, with no distinction being made between tenses and
>the other three dimensions except which is more important for the currant
>narrative.

Very interesting. This is very broadly similar to the structure of
loglan. I didn't just clone loglan, however; I remapped space and time
to vectors so that the quantified tense is so much more natural to deal
with; in addition we have the assurance that if it works in math, it
must work in language. Due to Stephen's input on this thread VXT now
also distinguishes time intervals from space intervals when desired.

>
>As with anything non-universal, the varriation of the world's languages is
>stunning.

Jack, what do you see as language universals, if any? (If you have time
for this question)

Jerry
>
>Sincerely,
> Jack Durst
>Sp...@sierra.net

0 new messages