[NGL-project] Re: NGL ngl text
Jack Durst
Date: Mon, 6 Mar 2000 02:02:50 -0800 (PST)
From: Gerald Koenig <j...@NETCOM.COM>
Subject: [NGL-project] Re: NGL ngl text
From: Gerald Koenig <j...@NETCOM.COM>
>From: Jack Durst <sp...@sierra.net>
>On Sun, 5 Mar 2000, Gerald Koenig wrote:
>> >From: Jack Durst <sp...@sierra.net>
>> >
>> >On Sat, 4 Mar 2000, Gerald Koenig wrote:
xxxxx
>> One of the things I like about the way you have designed the lang
>> is the abstract vocabulary, like 'ec and fan. At times though I want to
>> be specific. As I walked yesterday the sudden thunder was without any
>It's more a matter of dialect philosophy than anything else. Tokcir is
>strongly derivational: Derive as much as you possibly can from existing
>morphemes. Zumirtok is more into coining and borrowing. It's your choice
>how to make your dialect, but every lexical choice does need to be
>justified.
>
I really do agree that "every lexical choice does need to be
justified". It's a major quality issue for the language. I don't
regard myself as a major contributor to the general lexicon, I'm more
focused on the tense systems, in all their manifold ramifications. I do
have opinions on the lexicon. I'd like to express them again, I'd like
to hear yours and Steven's again, so that we and other contributors can
develop a "corporate culture" , however diverse.
I am not opposed to derivation per se. I only dislike it when it gets
unwieldy and opaque. I like it especially for infrequent concepts where
it is a kind of mnemonic. It seems good for modules. I wish we had a
handy reference of the frequency of words for use in new word
construction. I don't see any derivations as fully self-defining. I
think they need to be treated the same as any other class of word with
well defined meanings and procedures for adoption. I think a special
automatic adoption policy for them is a mistake.
I would appreciate a little tutorial from you on your preferences and
style for structuring derivations. Perhaps I would use them more if I
understood them better, and newcomers could apply your rules. I don't
know whether I will agree with them or not, but I would give them a
try. Julian once said he prefered de novo creation of vocabulary. I
hope Julian and Steven will speak for themselves on their philosophys.
We have to remember that many others will eventually contribute.
As to myself, my first criterion is to not exceed the length of English
unless it is an obsolete English word; thus I wouldn't mind if "sword"
had 4 syllables and were a derivative. I am constantly mindful of
trying to make NGL better than english in every way. I guess I am alone
in being open to adoption of some, especially obscure, natlang
vocabulary. I believe in a deep connection between sound and meaning,
and that is a criterion for me. I look to natlangs for that
connection. I look for contrast in the sound of words likely to be
confused. I have little concern for economy of lexicon size, I feel
unused words will disappear or become of no consequence. For me, using
VXT, NGL is a kind of "Spanglish" because it mixes a latin style
grammar and lexicon with an english one. I expect to see and use a well
mixed language, even though I am developing pure VXT at this time. So I
expect a lot of long derivations and a lot of short english words
coexisting. I expect that there will be unforseen benefits to a well
integrated language.
Just one other personal thing. One of the first words adopted under the
voting and seconding rules was my proposal, Jack, you seconded it,
<verbixgeflait>::- completely satisfied at all levels of being.
It has fallen through the cracks; Stephen, could you put it in your
wordlist?
Xap nif je verbixgeflait,
Jerry
------------------------------------------------------------------------
eLerts!
It’s easy. It’s fun. Best of all, it’s free.
http://click.egroups.com/1/2073/0/_/_/_/952336972/
------------------------------------------------------------------------
Jack Durst
> >> >On Sat, 4 Mar 2000, Gerald Koenig wrote:
> xxxxx
> >> One of the things I like about the way you have designed the lang
> >> is the abstract vocabulary, like 'ec and fan. At times though I want to
> >> be specific. As I walked yesterday the sudden thunder was without any
> >It's more a matter of dialect philosophy than anything else. Tokcir is
> >strongly derivational: Derive as much as you possibly can from existing
> >morphemes. Zumirtok is more into coining and borrowing. It's your choice
> >how to make your dialect, but every lexical choice does need to be
> >justified.
> I really do agree that "every lexical choice does need to be
> justified". It's a major quality issue for the language. I don't
> regard myself as a major contributor to the general lexicon, I'm more
> focused on the tense systems, in all their manifold ramifications. I do
And your contribution in that regard is by all accounts an excelent one.
VTT is the most through and logical tense system I've ever seen, though
I'm less comfortable with it than I am the other two.
> have opinions on the lexicon. I'd like to express them again, I'd like
> to hear yours and Steven's again, so that we and other contributors can
> develop a "corporate culture" , however diverse.
I understand, andagree completely.
> I am not opposed to derivation per se. I only dislike it when it gets
> unwieldy and opaque. I like it especially for infrequent concepts where
> it is a kind of mnemonic. It seems good for modules. I wish we had a
> handy reference of the frequency of words for use in new word
> construction. I don't see any derivations as fully self-defining. I
> think they need to be treated the same as any other class of word with
> well defined meanings and procedures for adoption. I think a special
> automatic adoption policy for them is a mistake.
On this position I would tend to disagree with you sharply. Words do not
need to be precisely defined in the vast majority of uses of language.
Our first stated goal is that NGL is to be a *human* language intended for
native speakers. Humans have reason, judgement, and common sense, they
are very tolerant of ambiguity and can almost always come up with a
reasonable interpretation of any meaning.
For example, english is very ambiguous in grammar, while having an
unusually precise lexicon. For example, the noun phrase "Italian leather
sofa" could mean a sofa made of italian leather, or a leather sofa made in
Italy, a sofa made in Italy of Italian Leather, or even a leather sofa in
the italian style. Each of these would be translated differently in NGL
{sofa loixoi' italig, sofa loixoi' italig itali (à Ital), sofa loixoi' à
Ital}. NGL prefers a precise grammar, but allows some leeway in
vocabulary.
It is both usefull and desireable that words do not have precise meanings.
It allows greater freedom in derivation, while still having a strong
chance of being interpreted exactly as you intend. If you need a more
precise word, you can paraphrase or further derive, but in the vast
majority of cases, precision is not nessicary and should not be forced on
the user. Having more general morphemes allows the morphemes to be of low
adverage length while still covering enough semantic space to allow the
usefull expression of ideas. Given that the morphemes are specific enough
to allow substantial derivation, a fairly high degree of precision can be
achieved, and almost all ideas can be expressed.
Doing it this way allows a trade-off between precision and speed, entirely
at the tokor's discression, it makes the language more powerfull by
allowing the exact degree of precision desired, but also more useable and
learnable, as one only need know roughly 700 morphemes to express one's
self competantly on virtually any subject at whatever degree of precision
one desires. Similarly, the tulor need only to recognize those same
morphemes and know some simple rules of derivation to understand whatever
is being said, in precisely the level of precision the tokor chose.
The goal of a language is *not* to convey all of the information
available, only to convey enough of the information to effectively get
your point across. High-precision words and constructions are a usefull
tool when required, and certainly have their place in the language, but
need not always be used.
> I would appreciate a little tutorial from you on your preferences and
> style for structuring derivations. Perhaps I would use them more if I
> understood them better, and newcomers could apply your rules. I don't
Derivation in Tokcir is a fundamental part of grammar to a greater extent
than any other dialect of the language (though exactly the same rules of
derivation are used in both Tokcir and Zumirtok.)
Derived words are always interpreted left-to-right, first the head, then
any general morphemes making the head more specific, then deriving
morphemes in order left-to-right, then {-(o)to/-(i)n(o)to} to make it
large or small, then the person if a verb {-o/a/e-}, then the number
(-(e)m/r/s), and finally the case {-om/-ac/-ad} if a noun or adjective,
or the mood/tense/aspect if TVS verb.
There are two basic rules:
1. For general morphemes, put the head first, and then tack on other
morphemes to clarify the head. {reda} "garden", {redabiru} "beergarden"
{kuaj} "vehicle", {kuajvod} "boat", {kuajvodinto} "ship"
To identify a head when thinking of English, look at the last element and
decide which element is the compound word a kind of, a "beergarden" is a
kind of garden, so garden goes first.
Inverses are treated as if {in`-} and the morpheme it attaches as if it
were a single morpheme. {inxòmcop} "to tear down a building". {in`-}
always attaches only to the morpheme it immediately preceedes.
Diminutives and their inverses can be incorporated within a word,
especially when they change the meaning of a word. The diminutive
attaches to everything before it in the same word, when placed at the end
of a word, it attaches to the whole word.
WARNING: Due to English influence in the language, placing a verb at the
end of a derivation from a noun is not reccommended, as the verb,
especially if common, will sometimes become the head. This is not
correct, but does occur both in Tokcir and Zumirtok.
2. There's a second class of morphemes which don't follow this rule, these
are the part-of-speach and pseudo-gender changers. Usually a word
inherits the part of speach and pseudo-gender (person / device / thing /
place / abstraction status) of its first element, this is usually
desireable, but it's also very derivationally usefull to be able to change
it.
Even though these things *should* rightly be the head, they're not
terribly meaningfull on their own, and it's much easier to read when the
first morpheme is the one which carries most of the meaning, so they go at
the end of the word instead. There are only a limited number of such
morphemes, however, and they can easily be learned by rote.
The part of speach changers are: -fe/-je, -i, -ig
The noun pseudo-gender changers are: (-st-) -or/-ir/-ar, -ci/-jan, -ka
{-fe} is used to turn any part of speach except a verb into a transitive
verb, or a verb into a causitive verb, {-je} is used to turn any part of
speach into an "is/has" verb. {-i} turns any part of speach into an
adjective, {-ig} turns any part of speach into an adverb. All of the
gender changers will turn any part of speach into a noun. {-or/-ir/-ar}
derive person/device/thing gender from the agent role of a verb, or from
the meaning of the {-je} form of a noun or adjective, the corresponding
-(s)t- form derives the same gender from the patient role. {-ci} derives
an abstract noun/adjective of quality from any part of speach, {-jan}
derives an abstract noun of process from a verb. {-ka} derives a term for
a place from any part of speach. {-kol} Derives a color adjective from
a noun or adjective.
The deriving morphemes can be combined in a series of successive
derivations to derive a wide varriety of words. For example:
ocod - medicine (the field)
ocodfe - to practice medicine
ocodi - medical
ocodig - medically
ocodor/ocodfeor - doctor
ocodorka - doctor's office
ocodir - medicine (the substance), also medical devices
ocodirka - pharmacy
ocodka - hospital
ocodfejan - the practice of medicine
Likewise: sek, seki, sekig, sekir, sekor, sekar, sekci, sekjan, seka...
fly, flying, airplane, pilot, (ne*), flight, flying, airport
* No english term, "thing which flies"
The derivations continue, as each of these words can further be derived by
a change of part of speach or pseudo-gender.
Combining the two, one can derive just about any word one needs, for
example: {kuajvodintofeorka} "Shipwright's shop"
To read a derived word, just go left-to-right deriving as you go along.
{kuaj} "vehicle", {kuajvod} "boat", {kuajvodinto} "ship", {kuajvodintofe}
"to make a ship", {kuajvodintofeor} "shipwright", {kuajvodintofeorka}
"Shipwright's shop"
WARNING: It is not advised to derive using a general morpheme from a word
which already ends in a part of speach or pseudo-gender changing morpheme.
These derivations do not occur in any text in either of the two dialects
with derivation, and the evolution of the language appears to have ruled
out this form. Though such words should be interpretable and meaningfull,
they tend to feel ungrammatical and interpretation of them is uncertain.
When a derivation has a special meaning (that is, of all the things a
derived word *could* reasonably mean, it is usually assigned a particular
meaning that you know), all further derivations from that word will also
have the same special meaning.
And that's basicly all you need to know to make and understand derived
words in Tokcir/Zumirtok, the rules of derivation are the same in both
dialects, and the Ogden set is shared between all dialects of NGL. On the
whole, Tokcir tends to be more derivational, though both make heavy use of
derivation.
NGL was not designed to be computer parseable, so some morphemes get
hidden (like {-ka} in {seka}) and sometimes there are ambiguities as to
morpheme boundaries. It pays off to know a significant portion of the
Ogden set by heart both when you're deriving and when reading derivations,
the best way to do this is by practice in reading and writing the
language. The more you read and write using derivations, the better you
get at recognizing the common morphemes. When in doubt, always assume
until proven otherwise: that a morpheme you recognize *is* that morpheme,
that it's a common morpheme instead of a rare or new one, and that the
most reasonable interpretation, given the context, is correct.
A partial corpus of Tokcir and Zumirtok writings is available in .rtf
format at: http://personalweb.sierra.net/~spynx/nglbook.rtf
The .rtf format can be read by word processors in most operating systems,
on Macintosh Simpletext will read .rtf, on Windows systems Write or
WordPad will read it, on UNIX I believe TeX and emacs may read it (though
I could be wrong there, check...) and I know WordPerfect does.
> know whether I will agree with them or not, but I would give them a
> try. Julian once said he prefered de novo creation of vocabulary. I
> hope Julian and Steven will speak for themselves on their philosophys.
> We have to remember that many others will eventually contribute.
Of course, and everyone is free to have their own philosophy. As I said,
derivation is my choice because of its profound benefits in terms of
flexibility and ease of learning.
> As to myself, my first criterion is to not exceed the length of English
> unless it is an obsolete English word; thus I wouldn't mind if "sword"
> had 4 syllables and were a derivative. I am constantly mindful of
> trying to make NGL better than english in every way. I guess I am alone
Trying to be better than English is as much of a natlang bias as directly
importing English constructions without a second thought. NGL should be
the best language we can build, and there are a lot of criteria on what
makes a good language.
One might choose ease of learning (interlinguas), or computer parseability
and logical organization (lojban) as a criterion. Fortunately for us, the
stated goals of this project narrow it down a bit. NGL is to be a human
language, which rules out computer use as a primary purpose, for native
speakers, which makes ease of learning a secondary purpose at best. It is
to be organized along the principles of linguistic universals, which rules
out logical organization (in logic, an even number of negatives make a
positive; in most languages, more negatives make a more and more emphatic
negative.)
I personally choose versatility while maintaining a minimal learning
burden and as close a correspondance with universals as possible. This
leads Tokcir to be a highly structured but minimalistic dialect. I cannot
speak for Stephen, but from what I've seen of his writing, his goal seems
to be to make the language as expressive as possible. Your criteria are
your own, of course. I chose to sacrifice adverage word length for
minimum morpheme length and a lower learning burden.
> in being open to adoption of some, especially obscure, natlang
> vocabulary. I believe in a deep connection between sound and meaning,
> and that is a criterion for me. I look to natlangs for that
> connection. I look for contrast in the sound of words likely to be
I have no objection to borrowing as a source of new morphemes, I simply
worry about the English and latinate influence on the language as a whole.
Some morphemes from a wide varriety of languages is fine, but not to many
from any one language, and leave some room for the a priori, too.
The look and sound of morphemes is important to me, too, and if the sound
of a word connotes the thing it represents, that's wonderfull, but the
most important thing is that they be NGL, and not some other language.
Let's face it, NGL was not designed to be a pretty language. The written
form of our common morphemes is butt-ugly. Sometimes this was intentional
(I picked {xuvuj} because it was the ugliest looking form on the random
list of valid wordforms I generated) but in the main, the morphemes were
initially chosen because they were mnemonic for me in some way, {badib}
because it sounds like someone singing scat, {exu} because it sounds like
the activity it denotes, {nemo} for captain Nemo the mad scientist from
20,000 leagues under the Sea; or because I thought it somehow suited the
meaning.
But, like it or not, there is a certain sound and look to NGL, and the
language gains a sense of coherence from that. {tixaw, zuoh, guva}
though all coined long after the ogden set are *classicly* NGL words both
in sound and orthography, and the language gains identity by their
presence.
Instinct in coining is just as good a way to touch that which is universal
as borrowing is...
> confused. I have little concern for economy of lexicon size, I feel
> unused words will disappear or become of no consequence. For me, using
For me, this is far different. You've seen my philosophy above, and to me
it's not only unnessicary to have a word which could just as easily be
derived as coined, it actually detracts from an important part of the
language, in that it produces yet another morpheme everyone must learn,
while adding nothing to the expressiveness of the language, and taking up
yet another slot of our dwindling space of valid short forms.
> VXT, NGL is a kind of "Spanglish" because it mixes a latin style
> grammar and lexicon with an english one. I expect to see and use a well
> mixed language, even though I am developing pure VXT at this time. So I
> expect a lot of long derivations and a lot of short english words
> coexisting. I expect that there will be unforseen benefits to a well
> integrated language.
Mixing in English to me is another horror, it raises the spectre of a
creating natlang bias in the language which unfairly favours one
set of languages over another. It blinds us to our own creativity in
creating the language, and to all of the other possibilities out there in
the world. I would have no opposition to drawing borrowings equally from
many obscure languages, but to favour one in such a blatant way riles me.
In terms of mnemonicity, a priori coining is just as easy, and far less
dangerous a way of obtaining it.
Sincerely,
Jack Durst
Sp...@sierra.net
[this posting written in Net English]
Jack Durst
Date: Fri, 10 Mar 2000 00:43:01 -0330
From: Stephen DeGrace <c72...@morgan.ucs.mun.ca>
Subject: Re: [NGL-project] Re: NGL ngl text
From: Stephen DeGrace <c72...@morgan.ucs.mun.ca>
On Mon, 6 Mar 2000 02:02:50 -0800 (PST), Gerald Koenig
<j...@NETCOM.COM> wrote:
>From: Gerald Koenig <j...@NETCOM.COM>
>
>
>
>>From: Jack Durst <sp...@sierra.net>
>>On Sun, 5 Mar 2000, Gerald Koenig wrote:
>>> >From: Jack Durst <sp...@sierra.net>
>>> >
>>> >On Sat, 4 Mar 2000, Gerald Koenig wrote:
>xxxxx
>>> One of the things I like about the way you have designed the lang
>>> is the abstract vocabulary, like 'ec and fan. At times though I want to
>>> be specific. As I walked yesterday the sudden thunder was without any
>
>>It's more a matter of dialect philosophy than anything else. Tokcir is
>>strongly derivational: Derive as much as you possibly can from existing
>>morphemes. Zumirtok is more into coining and borrowing. It's your choice
>>how to make your dialect, but every lexical choice does need to be
>>justified.
>>
>
>I really do agree that "every lexical choice does need to be
>justified". It's a major quality issue for the language. I don't
>regard myself as a major contributor to the general lexicon, I'm more
>focused on the tense systems, in all their manifold ramifications. I do
>have opinions on the lexicon. I'd like to express them again, I'd like
>to hear yours and Steven's again, so that we and other contributors can
>develop a "corporate culture" , however diverse.
To me feel is important. When I write, I like to try and hit the
"shade" of meaning "perfectly", which probably has much to do with my
preference for coining. I'm a bit perfectionistic, sometimes. Whereas
Jack seems to like to go with the flow and only add to the vocabulary
if forced, I have a more tempestuous outlook and seem to prefer to
just ram my way through to the meaning I want by coining :-). Both
approaches have merit, in my view, and both are complimentary. I work
at growing the supply of new monomorphemic words, and Jack keeps it
from getting out of control :-).
I guess if _I_ had to characterise the dynamic in here form my own
point of view, I'd say roughly that I'm interested in building a
bigger toolkit of monomorphemes through experimetation my experience
in composition as a tool to point out what might be good morphemes to
have, and Jack is interested in making sure that nothing gets in there
that's not justified.
>I am not opposed to derivation per se. I only dislike it when it gets
>unwieldy and opaque. I like it especially for infrequent concepts where
>it is a kind of mnemonic. It seems good for modules. I wish we had a
>handy reference of the frequency of words for use in new word
>construction. I don't see any derivations as fully self-defining. I
>think they need to be treated the same as any other class of word with
>well defined meanings and procedures for adoption. I think a special
>automatic adoption policy for them is a mistake.
The way I look at it, most derivations are not "automatically adopted"
as they are inherently of lesser status unless a specific effort is
made to give them official status. It is theoreticaly possible to make
a gargantuan number of derivations, and this informal and organic
system for dealing with most of them gives you a little leeway to
figure out what the "good" ones are, and keeps us from getting bogged
down in justifying every single derivation.
>I would appreciate a little tutorial from you on your preferences and
>style for structuring derivations. Perhaps I would use them more if I
>understood them better, and newcomers could apply your rules. I don't
>know whether I will agree with them or not, but I would give them a
>try. Julian once said he prefered de novo creation of vocabulary. I
>hope Julian and Steven will speak for themselves on their philosophys.
>We have to remember that many others will eventually contribute.
Well, derivations are head first. Basically backwards of the way
English constructs compound words. For me, there are some exceptions,
though. when I make a compound using a vector morpheme, which I do on
occasion, especially if the compound is a verb, I either park the
vector on the _front_ of the compound, or let it float out on its own
in the sentence as sort of a "separable prefix". As to mode of
vocabulary building... for me, highly intuitive. I have little faith
in scientific systems, certainly I am sure I cannot construct one
myself I would have faith in. But I have a developed intuition from
study of other languages and linguistics on an amateur basis, and if
it is not too immodest to say so, a certain talent in this area. I
trust this, coupled with some logical reasoning on a case-by-case
basis, to produce my coinings in such a way that the whole will work
together. The trust is based on not having any better way to do it
:-). I may or may not let myself be inspired by a word existing in
another language, and I try to carefully consider the semantic space I
want for the NGL word.
>As to myself, my first criterion is to not exceed the length of English
>unless it is an obsolete English word; thus I wouldn't mind if "sword"
>had 4 syllables and were a derivative. I am constantly mindful of
>trying to make NGL better than english in every way.
This is not a goal for me, I would not know how to measure it. I just
try and do my best, and we'll see how it all works out in the end. I
figure that on average it has basiclaly worked out although for me
size isn't everything ;-).
>I guess I am alone
>in being open to adoption of some, especially obscure, natlang
>vocabulary.
Well, I allow myself to be inspired by other languages, sometimes big
languages, sometimes little ones. The NGL species word I coined for
"human being", {duin}, for example, comes from a similar Irish Gaelic
word that means just that.
>I believe in a deep connection between sound and meaning,
>and that is a criterion for me. I look to natlangs for that
>connection. I look for contrast in the sound of words likely to be
>confused.
I have no opinion on that. I'm open to the idea. I certainly appeal to
an "inner intuition" for a word that has a sound that somehow "feels
right" when I coin. I think I'm appealing to an aesthetic sense more
than anything, though.
>I have little concern for economy of lexicon size, I feel
>unused words will disappear or become of no consequence.
I worry about that a little, actually. But I see your point and am not
too fussy about lexicon size, within reason.
[...]
>Just one other personal thing. One of the first words adopted under the
>voting and seconding rules was my proposal, Jack, you seconded it,
>
><verbixgeflait>::- completely satisfied at all levels of being.
>
>It has fallen through the cracks; Stephen, could you put it in your
>wordlist?
It is done, my friend :-).
Naesverig,
Stephen
------------------------------------------------------------------------
Shabang!com is the place to get your FREE eStore, Absolutely FREE
Forever. If you have any desires to sell your products or services
online, or you want to expand your customer base for FREE, Come check
out Shabang!com FREE eStores!
http://click.egroups.com/1/1299/0/_/_/_/952661485/
------------------------------------------------------------------------