Re: LUNATIC again

At 5:23 PM -0400 11/6/98, Logical Language Group wrote:
>David Durand:
>>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...
>Yet, I think most people can look at a conlang and decide whether it
>constitutes "encoded English".  I don't think that grammar changes such
>as aspect matter that much, unless they are USED differently.
If the features _don't exist_ in English, or are mandatory in the language
and optional (or only expressible by complex paraphrase in English) then
they are by definition used differently from English, since it does _not_
make the distinction usually.

Aspect is not the ideal case because it's pretty universal (more common
than tense) -- but also highly variable in how it's divided. Nonetheless,
introducing such distinctions _creates_ a difference of usage, since even
word-by-word translation hs to make such usage distinctions.


>>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.
>
>Yet it seems a perfectly sensible question to ask "Is he bald?"  The
>definition
>may not be worded in a way that tells the answer, but the person answering the
>question can.  All this shows us taht there are limits to definition-writing
>and not that people don't know what "bald" or "language" mean.
It also shows us that some kinds of definition are inherently bogus: like
your definition of language. It's one thing to list features or examples
for a fuzzy term, and then admit the use of judgement ind etermining how
that term applies. It's another thing to claim a fixed meaning that can
judge "this is a langauge" and this "isn't a language". There are many men
and women in partial states of hair loss where it's not easy to tell if
you'd call them bald or not -- or where you might even ask: "Do you mean
_really_ bald, or just _mostly_ bald?" or something of the sort.

I claim that language is like baldness, only more so, since it's
multidimensional, over things like number of L1/L@ users, fluency of
speakers, potential speakability (in terms of human grammatical
capabilities), compositionality of semantics, richeness of vocabulary,
applicability to daily life, etc.

Personally I think of a language as an abstract system of signs _capable_
of being used for communication. Just as we don't require a code to be
used, to tell if it can encode the messages we intend to send, we already
know enough to tell if a proposed system is _adequate_ to be spoken by
human beings.

That's exactly the question you had to answer in finalizing Lojban. You
certainly didn't have to reach a critical mass of speakers to know that.

> >A recent project of mine has a "narrative" mood. English has no such thing.
>Not in the sense of a gramamtical feature.  But English does have a
>"narrative style" that has a set of conventions.  Lojban is indebted to
>tense linguist/logician John parks-Clifford for pointing out the nature of
>"story time" as a tense convention that occurs in many languages.  See the
>tense chapter of the reference grammar for details.
I know about this phenomenon -- that's one reason for the grammatical
feature. But a style is a different thing from a mandatory grammatical
marking, and the implications for usage are very different as well. We
might disagree whether some English text was in a "narrative style." In my
as yet unnamed project, you would examine the morphology of the verb to
answer the same question.

In daily life, the choice of verb form would be a signficant factor in
telling someone about a past event.

>I think that when people first use a slang expression, the wordsdo have close
>association with a more standard equivalent, but then the metaphorical
>extensions start growing.  "Cool" does not mean what it did when I
>was a kid, when it
>could in most cases have been exactlyu replaced by a number of other words
>with no change ion meaning, in the typical slang usage.  If there were
>subtleties of meaning in the various slang words of approval, they were lost
>on me.
Maybe, I'm 37, and it's had a very complex set of meanings my whole
linguistic life. It seems to since before I was born, too, judging by what
I've read.


>This is fair.  However I did not intend in my posting to make my definition
>the issue.  It was rather more a statement of the criteria that a linguist
>might use to rule out relevance of a conlang.  That these criteria might
>be arguied against with hypothetical or evn real exceptions is not
>critical - the point is that they are some of the types of things that
>will be used to challenge
>a conlang seeking acceptance as an "interesting language".
Since you apply these criteria in a naive way to declare some things "not a
language," attacking the criteria is relevant. I think some things you rate
very highly are not that important. What I think is _very bogus,_ and leads
to misunderstanding, is your using a binary language/not language
distinction, where a multidimensional rating of characteristics is the
_only_ appropriate thing to do.

>In most every scientific field there are the occasional person, typically
>rejected by the mainstream, who comes  up with an important new idea which
>gets eventually accepted.  Famous patron is one way this acceptance comes
>about, but there are other ways.  But mostly I think success is
>determined by stubbornness coupled with learning to play the political game
>that is peculiar to the field.
It happens. As far as I can tell, _very_ rarely. What usually happens is
that such people are ignored, but may be noticed decades later when
external factors lead to the rediscovery of their ideas by others. Then
someone remembers that it's not _really_ new. Continental drift is my
canonical example here.

It's less likely when the proposal is an experimental paradigm and not a
single provable thesis, or technological innovation.

>>You set up a definition of
>>language (with which you were unjustly attacked) and used it to attack
>>--More--
>>artlang projects.
>
>To correct you, I specifically tried NOT to attack artlang projects.
>If someone says that they invented 5 artlangs last year, I will not challenge
>them or denigrate them in any way.  If they say they invented 5  "languages"
>last year, then I will presume that they are applying the word language in
>a way that I do not accept, and which I think that most linguists would
>not accept, even if the precisedefinition they would use can be argued about.
I don't see you, or the semi-mythical close-minded linguist, as the arbiter
of what the word language means.

I'm like Humpty Dumpty, perhaps. "When _I_ use a word it means what I want
it to mean!"

Seriously, in an artificial language discussion list, at least, I think the
rough definition you want to apply is the marginal one.

>each person has their own criteria for what constitutes "good art".  If it
>is an artlang, I will evaluate it as an artlang.  If ity claims to be a
>language,
>then I will evaluate it as a language.
I'm sorry, but the two things are _not_ separate. Sure, the Art may be good
or bad. But the claim to being a language seems separate. It could be bad
art, but a perfectly reasonable language (some relex of English, say).

It could be a fully independent language from English in semantics and
vocabulary, but have no speakers.

It could be Esperanto, and fit your definition of "language".

The unitary term language hides the interesting extra-esthetic factors:
syntactic/semantic closeness to English, vocabulary size and scope,
grammatical power, learnability by humans (Jacques Guy has posted some
wonderful grammar sketches of adequate communication systems that would be
inherently unspeakable by beings with our nueral wiring).

The question "Is it a language?" is a red herring, luring us into extended
argument to no purpose. For language constructors the definition of
language needs to be very broad, but also differentiated in ways for which
many linguists have little use.

This isn't sci.lang.

> I am merely being honest in saying that
>I evaluate things sifferently based on the label put on them or the claims
>made about them.  Calling something a language involkes a bunch of concepts
>that I associate with languages based on my own experiences, and I respond
>to what is called a language based on those.  calling something a work of art
>tells me to invoke artistic aesthetic principles first and other criteria as
>a lower priority.
For an artlang, the existence of a skeleton that is (at least in principle)
completable (often simply by more vocab.) is a critical factor. The
expressive power is an important factor. You can have expressively
incomplete artlangs (the ones that attempt to force particular ethical
distinctions, for instance), but that incompleteness is part of what they
are. I don't aspire to an active speakerbase. But I do aspire to some level
of speakability. As far as a speakerbase, I really just want one that I'll
be happy enough to compose things (however laboriously), so I can create
cool illuminated manuscripts of the texts.

>Like Matt and John discussing model airplanes, if you call your model an
>"airplane" then a test pilot might be chagrined to find out that he cannot
>pilot it.  If you call it a "model"  hemight nbe quite interested to see it.
>If you tell me that you built 5 airplanes last year, and your name is not
>Hughes, Boeing or Douglas (and this is not the 1930s), I will presume that
>you are not plannning on having these planes carry either passengers,
>freight or weaponry, and my interest might vary as a result.
Well, this is nothing to do with whether it's a "plane". In this case, the
question again becomes more of what the purpose is. The military is
developing planes the size and shape of "model planes". Fitted with
imaging, computer control, and radio transmission, these will serve a
multitude of practical functions on battlefields.

A model plane _can't_ fly. A real plane _can_ fly, even if no one ever
flies it.


>I am NOT trying to exlcude artlangers from the set of people who invent
>languages.  Far from it.  I am merely showing that more precise use of
>English can prevent misunderstandings and misevaluations of your work.
>People WILL respond differently to things that are called "art" than to
>things that purport by their label to be the "real thing" that the art is
>supposed to resemble or represent.
No, you said above that it's _not_ a language if it doesn't meet your
conditions. In fact you've repeated the fact. You are claiming that certain
patterns of usage (at any point in past time) are essential to "being a
language". For that matter, usage in the future could be relevant as well.
Was Esperanto _not_ a language until Zamenhof achieved a sufficient base of
speakers? And did his non-linguistic productions (written in Esperanto
before its speaker base was large enough) suddenly become linguistic
objects once it did reach that size? This seems bizarre. The same written
text was not linguistic before some social event was complete, but it was
linguistic after tha social event transpired. Furthermore, the meaning that
any person with an Esperanto grammar would have produced from that text,
did not in fact change.  And this magic property of "language-hood" will
remain for all future time, even if all knowledge of Esperanto were lost
except that text.

>>One word-glosses are a characteristic of expediency of documentation, I
>>think, not just of poverty of distinction.
>
>But it is impossible to tell, unless you actually see usages where the
>one-word
>glosses differ in meaning from the language they translate.  Alltogther too
>often, people have clearly failed to consider the problem of avoiding a
>relexification of English semantics.  It takes hard work to avoid distinctions
>that in many cases are subconcious.  maybe most conlangers (including
>artlangers)
>do this work.  But the posts of lexicons that I see on this list usually
>don't give any clue to this work.
This is true. My question is why assume that they _didn't_ do this. And, in
any case, it's still a different language if the grammar is different, even
if the word-concepts map one-to-one.

At least I would say so, because even if the volcabulary is similar or
identical in semantics, the meaning of complete utterances my be very
different, depending on what is grammaticalized. These things compound as
language forms become larger, as well.

Consider nonce-language "English-prime." It has 3 aspects
(perfect/imperfect/repetitive), no tense, and no number distinctions. Most
individual sentences will seem quite close in English and English-prime.
But complete narratives will diverge much more, since different strategies
will be needed to control temporal placement (more temporal adverbs in
English-prime). More explicit numerical phrases may be needed. The will be
more ambiguities in pronouns, now that singular and plural are not there,
so repetition of noun phrase will move from being bad style to being good
style.

The repetitive aspect removes the need for expressions such as X-ing and
X-ing, he keeps X-ing, etc. Some verbs may now take that aspect by default
(or rule) -- e.g. "rock" (as in back and forth).

The preservation of homonyms would be bad, but few conlangers do that, even
when the use the "lone gloss" style for their vocabularies.

Insofar as you claim that language is definable, I disagree. And I
definitely think that such a single-minded notion is out of place here,
where so many need to push the boundaries of language.


  --- 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                    \__________________________