Re: LUNATIC again

At 3:35 PM -0400 11/5/98, Logical Language Group wrote:
>David Durand:
>>The crucial reason most linguists don't want to hear about conlangs is that
>>they study "languages in the wild." Considering conlangs of any sort is
>>bizarre, because it's not what they do.
>
>Matt has essentially said this.  However, I HAVE had success in gaining
>interest of some linguists in Lojban.  And there are linguists who do work on
>"interlinguistics".  Some of the former are even mainstream linguists of
>note.  The latter is still a rather specialized arena of linguistics that is
>not too well-known.  But there has been at least one linguistics conference
>organized around the theme of constructed languages.
I think this is great. It's an idea that has to have some interest for any
linguist (though it clearly offends some on an emotional level, a reaction
that I don't understand). Then again, I don't understand athletic fandom,
so I'm clearly not in sync with the rest of the world.

[I think people feel about the Red Sox the way I feel about Apple Computer
Inc, minus the self-interest that I have in the continuing existence of the
corporation (so I can buy more stuff from them). I think this way of
understanding it rather highlights than diminishes the gap in world
views...]

>It will be a long battle to make conlang linguistics more "mainstream".
>Unfortunately, the first step in becoming mainstream is to not seem to be
>challenging the existing mainstream so that you are rejected from the get-go.
>This means that I have to assume some of the attitudes of an academic linguist
>towards the discipline, cater to the existing definitions while trying to
>evolve them in directions that will aid my/our cause.
Right. I understand that you have been attacked. I question whether _all_
linguists accept those definitions however. We've already seen differences
of opinion in this list (from a "professional linguist in training").

I think you are giving the definition a weight that it doesn't deserve
because you've been repeatedly hit over the head with it. That is to say
that it's a good blunt instrument, but in practice it's not central to the
reaction that you deplore.

>>I think definitions are really incidental: the sociology of linguists makes
>>unnatural languages marginal at best.
>And I intend to change that.  Indeed the whole rationale of the Loglan project
>requires it.
I suspect that those experiments may not be done for decades or more --
There are too many other things that linguists already have to do, and
Sapir-Whorf is _so_ unfashionable. Of course S-W is often oversimplified
into a straw man, and then demolished -- not helpful in attracting people
to the investigation of what will surely be a complex and subtle series of
experiments, with controversial and contested results if _any_ effect can
be shown.

>>I have read similar statements to the ones below in some linguistics books,
>>and they offend me there too...
>
>Well, it comes down to "who is the arbiter of the meaning of a word", the
>masses or the experts.
The point that I accidentally _de-_ emphasised (it appears later in this
message) is that the "experts" admit in those texts that their definitions
of "language" are inadequate. They usually shuffle the question off to
philosphy, and ignore such definitions in practice, because in practice the
sociology of the discipline tells them what is linguistics, and what is
valuable within it.

> So which definition of math wins?  In this case, the masses win.  If you are
>seekingf to publish in a math journal, the masses lose.
Again, the problem is not really one of definitions: the problem is one of
what problems are deemed worth addressing. S-W is widely viewed as
discredited and attractive only to cranks. Artificial languages are
curently marginal to the active programs of research (theoretical ones
might have some interest, but don't currentlyl descriptive and functional
linguists are not interested in the problems you propose to address at all).

Even if you were to find a definition (like Matt's) that supports the
"language status" of Lojban -- and I'm sure such can be found, the
experimental method and intent would be close to unfundable.

I think this is a shame, and I suspect that there's more validity in S-W
than many give it credit for, Although I think the case for its "strong"
variants is very weak.

>>I will attack these definitions as an example of why I think this kind of
>>definition is futile. Another good thing to read is the "what is language"
>>section of intro. Linguistics textbooks. They never have a satisfactory
>>definition, and they usually admit as much.
>
>But the people who write the books are the ones that review trhe grant
>proposals
>and control the education of the next generation of linguists.  Their opinions
>count more than yours does, or mine.  What they say is a language, is a
>language.  Until we change their minds by showing that something that doesn't
>fit their definition, can be shown to be USEFUL if included.  Since you
>use the example of sign language of the deaf, this is now taken as being
>"language" primarily because people before mefought from within the system
>to make it so, by showing that the study of sign language gave results
>consistent with and useful to the rest of the body of linguistic research.
Again, I think that the definition is not central. In fact, I think that an
accurate definition of what linguists actually study would inevitably admit
of many "artificialities" which would weaken the case significantly. Some
of these cases have already been cited (language academies, Hebrew revival,
for instance). I'd include Sanskrit, which changed radically , but _always_
in accordance with the venerated rules set down by Panini. In some sense
Sanskrit is a conlang. (There are also many places where Sanskrit grammar
was purified in odd ways due to faulty analyses, and these were then
codified into the rules of composition followed for centuries).

>>>1. A language is a means of communication.
>>
>>Presumably a system of signs that can be used to communicate. Otherwise
>>Latin, or any language not in active use is not a language because it is no
>>longer the means for anyone to communicate if it is not actually used.
>
>I guess I have internalized the Lojban tense system too much.  I did not
>intend the present tense by using "is".  Future and past tense also can
>be taken (future being a fuzzy truth basedon the potential for it
>to become reality).  Latin was a language, so it is a language.
So if anything was ever a language then it is one forever.

So all I need to do is compose an email to someone in one of my langauges
and have them understand it, and then it's a language?

It's not just a tense problem.

>>So all dead languages are not languages, but "systems of signs that once
>>were languages". This is even more bizarrely true for dying languages,
>>since the next to last speaker takes the language with her when she dies
>>(sooner, actually, if she doesn't talk to the last speaker until the moment
>>of death). For some period of time it's in a limbo state, where it's not a
>>language, then it is used to communicate, so it's a language for a while,
>>then it's not used again, and it's not a language. If you say it's a
>>language between all the times that it's used, then it stops being a
>>language the last time the two talk, but you can't tell until one of the
>>two speakers dies.
>
>Ya know, I grow to appreciate the Lojban tense system every line you write %^)
>You are in effect questioning whether something must be continuous to be
>true.  Carrying your reasoning to its extreme, at a moment in which all
>speakers of English were silent, English would  not be a
>language.  And  at the moment between inhaling and exhaling, you are
>not breathing.
Exactly. My point was that there are a host of problems in the notion of
"being a langauge" The next tack to take is to say something about a line
of transmission. There are several ways to attack that, by counterexamples
to any definition of transmission that might be offered. Then once those
are set straight, there is sure to be some "non-language" that slips in
somehow, and that can be used to attach the definition again. These are all
based on classic conundrums of Anglo-American analytic philosophy. I'm very
sure that anyone attempting to defend such a definition can be worn down,
since they're all _unsolved_ classic problems. The linguist at some point
will give up, because she's a linguist and not a philosopher, throw up her
hands and say "enough of this logic chopping -- it's just not a _real_
language."

>>>4. Linguists also distinguish between "languages" and "codes", where the
>>>latter may differ in surface form from a parent language, but where the
>>>semantics of the "words" in the code are essentially unchanged from the
>>>parent language.
>>The intent of this is obvious, but the issue of whether semantics is
>>preserved between differing languages is a very open one.
>
>Open to whom?  Not to me.  I know too many words in Russian that do
>NOT mean what their counterparts mean in English, and too many that seem to
>mean the same up to the point where they mean something entire different.
Sorry. I was too brief. I meant the notion that there are  concepts in one
language that cannot be expressed in another language at all. Many make
this claim, but several factors argue against it: the reality of effective
translation, and the ability of humans to adapt and learn (and use the
compositionality of their own language to extend it in any direction they
find themselves needing).

Most linguists believe that the core semantics of human beings are
universal (this is a core part of the disbelief of the S-W hypothesis), and
thus the difference between a language and a code must be a matter of the
complexity of the translation function.

This is a classical problem of the continuum, and can be attacked by first
establishing a precise definition of "complexity" and then arguing that
process required to produce utterances in a language must have a specific
computational complexity relative to the native language of a speaker. Then
there's the question of what level constitutes enough to be a "language"
instead of a code. The only way to keep the definition under such an attack
makes it look too ridiculous to be credible...

The classic linguistic problem of the continuum is mutual intelligibility
of dialects. There simply _is_ no answer as to "how many" languages or
dialects there are if the gradation is slow enough.

The "intro to skepticism" version of the argument takes a man with a full
head of hair and removes one hair. Is he bald? Of course not. But if you
keep doing that, eventually he will be bald. When? Logically it seems that
there must be some point, but we all know that it's not really a very
sensible question.

>>Since most artlangs define grammatical features (like aspect, tense,
>>gender) that are different from the artist's native language: the semantics
>>argument is actually the weakest of the lot, even if you believe that we
>>know very much about semantics. (I think we speculate a lot, but _know_
>>little).
>
>They may define these things, but are the results of usingthem unlike the
>native language?  It is possible to map Russian perfective tenses to
>English usage (but not English tenses),and even easier to map them to Lojban
>tenses.  But the mapping is imperfect WHEN FACED WITH ACTUAL USAGE.  Without
>the usage, you could map things algorithmically.
I'm sorry, but if I add distinctions to my language that aren't in English,
or remove distinctions that are _mandatory_ in English, no pattern of usage
that I make will be a simple encoding of English sentences. Even if I
translate word-for-word from English, I have to decide things about the
meaning of the English that were ambiguous in the original (because
otherwise I can't finish the sentence). I also have to eliminate
distinctions that were in the English (number, for instance). If that makes
the result unclear, I can either leave it be (as a feature of my language),
or use some other way to make the utterance precise.

A recent project of mine has a "narrative" mood. English has no such thing.
It's easy to decide that any extended utterance starting "Once upon a time"
would use the narrative mood, but most other situations require a conscious
decision. I think I'd use it at the end of the day when talking with my
family over dinner, but not when referring to a single incident in the
past. But this is a decision that can't be made based on the English.

That's what I meant by a different feature. Sure, it may reflect my
preoccupations or culture, but I don't see that as proof that it's encoded
English.

>>This is essentially specialized vocabulary with meanings hard to express in
>>the rest of the language.
>
>Not necessarily.  Slang words are a cant, usually created by the young people
>or by a  minority, in order to set themselves apart from others.  the words
>may have identical meaning to someother word, but a different word is used
>merely to show differentness.
Define "Cool.".

You could write a dissertation on it, I suspect.

At one time I could have written a dissertation on the differences between
bakeage, fryage, macro-fryage, toasty around the edges, and roast-a-roma --
all mental states of some ineffability, and great importance to me and my
friends at the time.

There were also metaphorical uses of these terms, as in "So she left, and I
fried!" (abandonment by girlfriend causes intense and lasting distress).

"It's time to go."
"let's hurl."
and "Bwana, I need to achieve macro-hurlage"

all signify somewhat differently in ways that were transparent to me (and 2
or 3 other people) but require long paraphrase in regular English.

I don't think these are synonyms for English words, but they were certainly
a personal slang. I think slang refers to a social phenomenon surrouding a
specialized lexicon, not to a fundamental feature or type of language.

>>I agree that some closed-minded professional linguists accept these
>>definitions or something like them...
>
>Thereby conceding my case.
well, in a limited way. I don't think that these definitions are the core
of your problem with the profession: the social status of outsiders to the
academy is one thing, the attachment to an issue that the profession
believes (perhaps wrongly) to be completely settled, the unusual nature of
the proposed method of answering your question. These are your real
problems. The "langauge" argument is just a convenient way to dump on you
so you will go away.

>>I deny that _these_ definitions hold
>>much water,
>
>They hold all the water necessary IN ACADEMIA.  THEY have the power, not you
>and not me.  Unlike artlangs, languages like Lojban need some sort of
>acceptance in order to achieve their purpose.  And the people who adjudicate
>acceptance are the ones whose definitions are all-determining.
I don't think so. I've never seen a definition of "language" that wasn't
explicitly marked as incomplete and provisional. (Eh... maybe bloomfield or
Gleason had one...). I think that they do hold the power, but the
definitions aren't at issue. You don't have the credentials (and perhaps
the background in Linguistics) that indicate that you're worth taking
seriously. Even if you did, you are far enough from the mainstream in
several areas, that you'd probably need a famous patron of you wanted to be
taken seriously -- otherwise instead of being dismissed out of hand as a
nut (_I'm not saying you are, just showing th reasoning_) -- you'd be
dismissed as a nut with a PhD in Linguistics.

If you kept quiet about it (except maybe in the bar after the conference)
probably no-one would hold it against you.

>Sure, all definitions have holes.  Among other things, I have read some
>of the theory of lexicography as a necessary step to learning how to write a
>proper dictionary.  That does not stop us from writing definitions.
right. In particular a definition of what language "really is", is a
philosophical definition. That's why I'm so sure that it can't hold up.

>>I'm not so sure: linguists study many things, but certainly central to
>>linguistics is the study of the grammar and semantics of human
>>communication. Whether that communication is potential (as with Tokana) or
>>actual and ongoing, as with English, or historical, as with Tonkawa (an
>>extinct native American language), they are all languages.
>
>The fact that Matt, in another post, has sais that he cannot use his
>Linguistics dept, Web site to put up his Tokana pages shows the falsity
>of the claim that all "languages" are equal.  I am sure that a 160 page
>study of any natlang would be far more acceptable use of his Web space.
>When this changes, I will accept your statement most gladly.
My guess is that his department can't afford to offer that much web space
to their students for _any_ purpose. They may be rich seeming to you, but
academia has a pretty small pie to share -- that's why academics are so
vicious in defending their share of it!

>>People can easily live within a vocabulary of a few thousand words, so that
>>even a perfectly respectable level of completeness is open to a conlanger.
>"People" can, but it is arguable that adults living in a 20th century
>society and possessing something akin to a college education cannot.  I
>have a
>vocabulary of a few thousand words of Russian.  I can talk somewhat fluently
>to a 6 year old (subject to some topical constraints).  I cannot even start
>to talk to a Russian adult about any topic I think that person could be
>interested in without resort to a dictionary or charades.  My situation
>with Lojban is somewhat similar - approxiamtely the same vocabulary size,
>except that
>there are no kids to talk to, and I have to coin new words on the fly ( a
>decidely non fluent activity) in order to converse with adults.
Right, but none of those "modern society" arguments has to do with the
definition of a "language" -- only a language for intellectual
conversations in modern society. You set up some ground rules for what a
language is, but these are different rules. You set up a definition of
language (with which you were unjustly attacked) and used it to attack
artlang projects. Now you're instead applying _your_ personal goals for
your language to argue that artlangs aren't real languages.

>I don't think it is accidental.  A complex society with intellectually
>active members needs a far larger lexicon.  If your conlang is intended for
>a real or fictional "primitive" tribe, a small lexicon might work.  If they
>have space travel, I doubt it.
As you say below, this is your goal, not mine. If you add your goal to the
definition of language, most languages of current and past human history
become some other things (incomplete communicative systems, or something).

>>Many of your points have to do with practical efficacy and vocabulary size.
>
>Of course.  I am not working on an artlang.  Practicality is the be-all and
>end-all of most everything I do.  Those who aren't trying to be practical
>are exempt from my comments and any real or implied criticism.  they are doing
>something else.  More power to them if they are having fun at it.
Right, but your claim was that they weren't "languages." And that's not the
same as saying they're languages that _don't meet my needs_.
The latter is a reasonable claim, I would argue.

I don't think that I've every created a language complete enough that I
would defend it. I think Matt certainly has, and some others here have as
well. Some are richer than others in different ways (vocabulary, suyntax,
literature), but there are quite a lot of languages that lack only
vocabulary to be "complete" in your sense.

Most of the artlangs I've seen certainly are not relexes, or anything close.

One word-glosses are a characteristic of expediency of documentation, I
think, not just of poverty of distinction.

  -- David
_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________