[YG Conlang Archives] > [ceqli group] > messages [Date Index] [Thread Index] >
on 2/26/02 2:13 PM, Mike Wright at darwin@hidden.email wrote:
> Rex May - Baloo wrote:
>>
>> Three languages impress the dickens out of me Ð Esperanto, Loglan, and
>> Mandarin. I want Txeqli to have the best characteristics of each.
>>
>> Esperanto. I want Tx to be able to build words in much the same way that Eo
>> does, from morphemes tacked together in a systematic way.
>
> I'll try to get through all the new (to me) info on the Web, but for
> now, I'll just ask questions to try to get up to speed on all that's
> happened. (Also, I'm quite ignorant about conlangs in general.)
>
> (I'll try to be consistent in saying "Mandarin" when referring to the
> standard, Putonghua/Guoyu, and only use "Chinese" when talking about
> the language family or the writing system.)
>
> I've been doing more reading on linguistics--especially regarding
> Chinese languages--since the last time I looked at Txeqli. One of the
> things I see about Mandarin is that there are some fairly universal
> structural principles. The ones that seem most basic are:
>
> 1) SVO: subject-verb-object word order is the sentence framework on
> which everything else is hung.
Same with Tx. Difference is that O can move around to make it OSV, and then
V can do likewise to VOS, but the order can never be reversed.
Go xaw zi. I see you. Zi go xaw. You, I see. Xaw zi go. See you, I do.
>
> 2) modifier-head: Modifying morphemes/words/phrases precede the thing
> modified. This principle is strongest in the construction of
> compounds, more variable in the construction of sentences. Noun
> modifiers precede nouns, verb modifiers precede verbs, and so on.
Same in Tx, with unmarked modification, like
Bon kan. good dog.
Which can, for clarity, be
Bon sa kan.
But Tx also has the option of moving the modifier (word or phrase) to the
end with 'hu'.
Kan hu bon. Good dog.
And I see no reason why verbs can't work the same way.
Kan kwai kom. Dog eats fast.
Kan kom hu kwai. Dog eats fast.
Normal will be modifier first, with or without sa. But the hu-form is
available if handier.
>
> Like all natural languages, Mandarin is full of exceptions--some real
> and some only apparent.
>
>> Loglan. I want Tx to have a completely straightforward phonology, like
>> Loglan did before they started messing around with it too much.
>
> Here's a question I asked back in February of 1997: Is Txeqli intended
> to have a native accent?
Hm. So far it's got a Southern Indiana one. I have no idea what to say
here. My first reaction is to say to emulate Italian. Does that make
sense?
>
> Also, have you made final decisions on:
>
> 1) permissible consonant clusters
I've been going by instinct so far. All voiced or all unvoiced, or unvoiced
prededing voiced. Worst clusters I've come up with is Bdomen (abdomen) and
Kfey (coffee)
>
> 2) permissible vowel clusters (diphthongs, triphthongs, etc.)
Anything is permissible here, I think. Each vowel makes a syllable. So a
word like pia is pronounced PEE-ah.
>
> 3) glides (on-glides and off-glides) vs. full vowels
Again, AEIOU are the only vowels and they always make syllables. So both on
and off glides have to be y or w. (I changed that. I liked the look of
'hau', for example, but consistency demands 'haw'.)
>
> 4) syllable structure
Seat of the pants here. I can't think of any potential Tx word where
there'd be a problem deciding. In a string, a syllable of course begins
with the first consonant cluster and ends with the following vowel, unless
you get a word like karlo, when there are two weaks together.
>
> 5) word stress patterns
Penultimate within a morpheme, but weak stress. baNAna. In a compound,
morphs keep their internal stresses. GER-man-pe.
>
> 6) phrase/sentence stress patterns
Not a clue on this yet.
>
>> I also
>> want it to have the ability to be completely unambiguous grammarwise.
>
> I'm not sure "completely" is a real possibility. If by "grammar", you
> mean "syntax", I think that you should look at the possibility of
> reducing ambiguity through the phonology--especially word-stress patterns.
Well, what I mean is that I want it to have the potential to be as
unambiguous as Loglan. Preciseness when necessary. Like the sort of
amiguous phrase. Go sa hon. My book. It's ambiguous in Tx and English,
because the 'my' can mean lots of things. It can be disambiguated in Tx
quite straightforwardly, thus.
Go ten sa hon. The book I have.
Go skri sa han. The book I wrote.
Tem go sa hon. the book about me.
>
>> Mandarin. I want Tx to have the terseness of Mandarin as well as its
>> extreme simplicity.
>
> Although Mandarin has little apparent morphology, its syntax is far
> from simple. Y.R. Chao wrote 819 pages on the subject in _A Grammar of
> Spoken Chinese_, and Li and Thompson then felt compelled to spend
> another 672 pages in _Mandarin Chinese: A Functional Reference Grammar_.
I'm going thru Li and Thompson as we speak. Yes, 'simple' isn't the word.
I like the way formal mandarin overlaps colloquial english, leaving out
unnecessary words, etc. If you look at
http://www.geocities.com/Athens/Acropolis/7429/chap2.html
on the old site (spelling, etc. is imprecise) you'll see that I was
attempting to see if I could be as parsimonious with words in Tx as was the
case in the Teach Yourself Chinese book.
Now, most of the sentences in that dialog could confuse a computer, I
suppose, so many of them can be disambiguated if necessary.
>
> To me, there are two big advantages for learners of Mandarin:
>
> 1) simple morphology: Mandarin tends to use separable affixes and
> grammatical particles, rather than making them inherent parts of words.
Loglan imitates this virtue, and so, I hope, does Tx.
>
> 2) optional syntactic categories: Mandarin permits the omission of
> many syntactic categories that are mandatory in many languages:
> gender, person, number, tense, aspect, definiteness. The implication
> here is that there is no default setting for any of these. They may
> truly be omitted. When the speaker decides that there is a need to
> express any of these categories, he/she can do so by adding lexical
> items ("yesterday", "male", "many", "one"), affixes (like the pronoun
> pluralizer, <men>), or particles (<le>, <zhe>, <ne>, etc.).
Yes, yes. Just what I'm getting at. I want Tx to be even _more_ that way
than Mandarin is.
>
> The simple morphology makes rules more universal, in contrast, for
> example, to all the irregular verb forms that occur in inflected languages.
>
> The lack of mandatory syntactic categories means that the user can
> learn a sort of all-purpose syntax in the beginning, and can put off
> learning how to express more specific ideas until they are required by
> the course of study.
>
> While Mandarin is full of the messiness of any natural language, these
> basic principles can probably be applied in a more regular way in a
> language like Txeqli.
>
>> Now, we can't have both unambiguity and simplicity. On the other hand, why
>> not? Unambiguity can be _optional_! Here's my cunning plan.
>>
>> It will be quite possible to speak Tx like Mandarin or various kinds of
>> pidgin.
>>
>> Kyu zi ten hon. Question you have book.
>> Ten. Have.
>>
>> Now, theoretically, ten and hon could be construed as a compound in speech
>> or by a computer listening, at least. So that possible ambiguity can be
>> eliminated by the t-word 'te,' which simply means that a noun follows and
>> isn't combined with anything before it.
>
> Another approach could be to have word-stress patterns for compound
> words. This could be revealed in the orthography by writing words
> without spacing. There might also be an argument for having a set of
> fundamental lexical compounds, with a particular type of stress and
> written as a single word, vs. more user-created compounds, with a
> different type of stress and written using some kind of hyphenation.
Hm. Yes, compounds shd be written as single words.
Dape DAH-pe. Giant
but
da pe DAH PE Big person
or
dasa pe DAH-sa PE Big person.
I don't know if people will use sa all the time or not. I find that I tend
to. It was of course based on Mandarin 'de.'
What with the self-segregating morphology, it would never be necessary to
hypenate so that readers could pick the morphs out. I guess a 'Txeqli
Academy', this list, that is, could create the initial compounds and admit
new 'official' ones as they'r thought up. Maybe hyphenating would be the
best way to propose new ones.
>
> One of the big areas of discussion on sci.lang over the years has been
> over how to define what a "word" is in a language like Mandarin, where
> the traditional orthography shows only syllables. Y.R. Chao, John
> DeFrancis, and dozens of other linguists have done a lot of work on
> this problem, and there is still a lot of disagreement. However, the
> latest work that I'm aware of shows that stress patterns play a big
> part in clarifying word structure in Mandarin and many other Chinese
> languages. The most obvious is the occurrence of the "neutral tone" in
> the second element of many two-syllable compounds.
I'm afraid I'm relying on natural English patterns in showing word
boundaries in speech. I'm not enough of a linguist to describe them.
But a 'word' in Tx is one morpheme or a combination of two or more morphemes
that has a definition that is more than the sum of its parts, like 'dape'.
Or so I think.
>
> Frankly, if the phonological system is too permissive, you are giving
> up a very useful element of the grammar.
>
> [...]
>> Another danger is the combination of the subject with the following verb:
>>
>> Kan kom te karn. Dog eat (noun) meat.
>
> Here's a place where stress patterns might be used to disambiguate the
> structure.
>
>> Well, we could do the parentheses thing again:
>>
>> Te kan bete kom te karn.
>>
>> Or, we can borrow a French trick and use the 3d person pronoun:
>>
>> Kan, da kom te karn. Dog, he eats (noun) meat.
>
> Here are some ideas from Mandarin that all fit the SVO/modifier-head syntax:
> (For simplicity, I'll stick with one verb form and one noun form in
> the English.)
>
> Kan kom karn.
> Dog eat meat.
> The dog eats the meat.
>
> kan kom sa karn
> dog eat GEN meat (where GEN is the genitive particle)
> the meat eaten by the dog
>
> kom karn sa kan
> eat meat GEN dog
> meat-eating dog/the dog that is eats the meat
>
> kan sa kom karn
> dog GEN eat meat
> the meat-eating of the dog
>
> kom kan sa karn
> eat dog GEN meat
> eating the meat of the dog
Yes, this is exactly what I want Tx to be doing. This last example is
probably one that would call for hu to clarify itself
kom karn hu kan.
>
> These are all phrases or sentences.
>
> Compounds might be distinguished by the use of stress:
>
> kom karn
> /"kom 'karn/ (primary stress on object of verb)
> eat meat
>
> komkarn
> /'kom "karn/ (primary stress on verb)
> carnivorous
>
> This could lead to:
>
> kom kan sa dir
> eat meat GEN animal
> animal that is eating meat
>
> komkan sa dir
> carnivorous animal
I think there probably would be little difference between the last two, and
I would tend to say that
Do kom karn sa dir. would be how I'd do the first one.
>
> komkandir (/"kom kan 'dir/?)
> carnivore
>
> Hope I haven't written up a lot of stuff that's already been
> thoroughly discussed and thought over.
Some of it has been thought about by me, at least, and it's very gratifying
to see that Tx has been well-enough described that you can run with it like
this.
One difference is that I've tended to tack sa to the end of the word,
as in komkarnsa dir, which may be a bad idea, as it tends to make me want to
stress 'karn'.
Gad. It's bedtime. Mike, you've really invigorated me and this whole
group!
--
>PLEASE NOTE MY NEW E-MAIL ADDRESS: rmay@hidden.email
> Rex F. May (Baloo)
> Daily cartoon at: http://www.cnsnews.com/cartoon/baloo.asp
> Buy my book at: http://www.kiva.net/~jonabook/gdummy.htm
> Language site at: http://www.geocities.com/ceqli/Uploadexp.htm
>Discuss my auxiliary language at http://groups.yahoo.com/group/txeqli/