Re: ANNOUNCE: My new conlang S11
|From:||H. S. Teoh <hsteoh@...>|
|Date:||Tuesday, March 8, 2005, 1:09|
On Sun, Mar 06, 2005 at 11:47:53PM +0100, Henrik Theiling wrote:
> "H. S. Teoh" <hsteoh@...> writes:[...]
> > It occurred to me, though, that with suitable noun marking, this might
> > not be that much of a problem. ...
> Yeah, maybe, but I wanted a language with only one word class and
> *only* syntax rules as structure. I introduced some prosody rules
> later to make the tree-structure clear, but I'm not so sure whether I
> like that. It's marked preliminary.
Hmm, OK. That's an interesting design goal. :-)
> > You could even use something really basic like have conditional
> > sandhi, that applies a mutation ...
> Again, a design goal: no mutation in Tyl Sjok. Pure isolation.
> > hausi fei sei -> hausi fisei ['hawsi fi,sej]
> Ah, nice!
> I actually like mutation and sandhi, but not in Tyl Sjok. :-)
> > Even with modern Chinese, sometimes native speakers' attempt to
> > describe it leaves foreigners feeling like there really are no rules,
> > it's just a matter of interpretation. :-P
> A misinterpretation, of course. :-) English and Germans and probably
> anyone might tell non-native speakers that there are no 'rule'. 'You
> can just say what you think.' Or something like that... That is,
> of course, not really helping. :-)
Indeed. :-) "Say what you think" has the hidden assumption that you
already have a mental grammar that matches the language. (Although,
that is an interesting thought... does this imply that you can't think
if you don't know any language??)
> > > You arrive at *exactly* the same structure above (including word
> > > order!), and at least for the motion verbs, the translation is very
> > > similar. Fukhian does have verbs for 'to kick', though, (and uses
> > > them) so most sentences would have a much more literal sense:
> > [...]
> > Cool. Another case of acadebism. :-)
> b? You don't mean 'better', do you?
Yep. Another Conlang's Already Dunnit Except Better. :-)
> I actually think that the Fukhian system, compared to your
> interpretation, is less elegant, since it has quite a restricted
> meaning, although the cases are used more generally with other
> Still, we *do* have similar ideals in conlangs. :-)
> Speaking of discoveries: when translating the Fukhian grammar from
> German into English, I was surprised to see that by my current
> analysis, Fukhian does not have verbs. Just nouns and adjectives. Or
> better: nouns and verbs syntactically behave so similarly that, when
> not knowing the meaning, it is impossible to tell from a sentence
> whether a stem is a verb stem or a noun stem. That's because:[...]
Interesting! I also noticed that at least at the lexical level, Tatari
Faran word categories may not be as clear cut as I make them out to
be. For example, a sentence with a monovalent verb looks like:
N <case> V COMPL
where <case> is a case clitic. An adjectival sentence looks like:
N <case> Adj COMPL
and a statement of equivalence ("X is Y") looks like:
N <case> N COMPL (COMPL here == "ai")
Now, if you saw a sentence in the following form:
N <case> X COMPL
where X is an unknown word, you have no way, at the lexical level, of
telling whether X is a verb, an adjective, or a noun! (Normally NPs
are overtly indicated by the presence of the trailing case particle;
but case particles are omitted from the predicate in "X is Y"
As I thought about this, it occurred to me that the reason English can
verb its nouns and noun its verbs so easily is because the (more or
less) fixed word order makes it much easier to tell them apart. For
example, take a quote from Calvin & Hobbes:
Verbing weirds language.
Even though the normal meaning of "weirds" is a plural noun, its
position in this sentence forces its interpretation as a verb. This is
quite different from the Tatari Faran case where the analogous word
position is shared alike by verbs, adjectives, and nouns.
> When I discovered this, I was a bit surprised, but I found no trace of
> a violation of this discovery. Very interesting, since I did not
> notice when I worked on Fukhian grammar. I always thought it had a
> distinction, but in fact, lexically, it has not.
I'm realizing the same thing. I had always tended to design grammar
from a purely abstract viewpoint: the verb goes here, the noun goes
there, etc.. Until recently, I've never thought about it from a more
"black box" point of view, where given a bunch of words, how would one
go about deciding which word is in what category, esp. if one's
vocabulary happens not to include some of the words present.
> > > 'I am tired.'
> > >
> > > This is a state, so what concept moves?
> > Ebisédian does not use verbs for states. ...
> Ah! That's logical. :-)
Yeah, this point actually stumbled me for quite some time when I was
designing Ebisédian grammar. For a long time I was so obsessed with
how to handle verbs, that I completely forgot about stative
statements. When I suddenly realized I needed a way to express those
as well, it took me a while to try to rationalize it with the
verb-centric system that I had developed. (This is one of the reasons
stative sentences in Ebisédian are so strange---they were patched on
after the fact, so some ugly kludges had to be made.)
> > forest(loc). Here, of course, the usage is very idiomatic: the literal
> > reading of this sentence is "I am inside fatigue".
> Hmm, is the locative restricted to 'inside'? No, right? Just general
> location, I assume.
Correct. Keep in mind, though, that it's not the locative itself that
gives the meaning of "inside", but the *combination* of the locative
with the conveyant. This is actually so that it is consistent with the
verbal paradigm for the 5 noun cases. Remember the metaphoric model of
the conveyant noun moving from the originative to the receptive, and
passing through the locative. Well, since there is no verb to effect
any movement, that means the conveyant noun is simply sitting at the
position of the locative. :-)
> Reading it like 'I am with fatigue' (with=accompanied by, near,
> located at, but not necessarily 'inside'), it makes perfect sense.
Yeah, that's the intent.
> Actually, Finnish expresses 'to have' with the normal 'essive'
> copula 'olla' + allative (one of the two locatives the language
> has), so 'I have fatigue' is perfect. Further, Russian uses a
> 'There is on me ...' construction for 'to have', which comes very
> close. And many German dialects use 'to have' for 'to be cold',
> which is also an expression of internal state.
Ebisédian uses another (some may consider 'crazy') way of expressing
position: the combination of the conveyant (the thing being owned)
with the receptive (the owner). This is actually adapted from Greek,
where the dative case is used for the owner. The receptive case does
behave somewhat like a dative. One way to understand this is that two
people are sitting at a table deciding to whom object X on the table
belongs. The result of the decision would be the transferring of X to
the owner, so the owner would be in the receptive. Hence the
> > Like I said, this system, although it does somehow "make sense" in
> > its own way, is too ambiguous for my tastes, so I've abandoned it in
> > Tatari Faran.
> Hmm, I still like it and find it very functional.
It's certainly functional, just rather hard to become fluent in. :-)
Fluency, after all, is one of my goals in conlanging... even if I
never actually attain to it. :-P
> Does Tatari Faran introduce verbs to handle this instead?
Yes, Tatari Faran is less pedantic about not representing state with
verbs; but Tatari Faran also has true adjectives, so things are a
little different there. The verb used to express "to have" is actually
a verb meaning "to acquire": _kuini ... dakat_. The acquirer is in the
receptive case, and the person acquired from is in the originative.
However, if there is no originative NP, the meaning becomes "to have".
fia nei kuini bibi sei bunari kei dakat.
Fia RCP acquire doll CVY woman ORG COMPL
Fia acquires a doll from the woman.
fia nei kuini bibi sei dakat.
Fia RCP acquire doll CVY COMPL
Fia acquires a doll. / Fia has a doll.
> > This means that "what" in the English question "what happened" can
> > serve both as an interrogative noun and an interrogative verb. You
> > can answer with either a noun or a verb equally validly.
> !! Impossible! I don't dare to say, but I still do: Fukhian's
> interrogative noun stem for 'which?'/'what?' can be used as a verb and
> means 'do to what?'. Thus there are questions like 'you whatted?'
> Just like in you lang! ~&->
> > Another interesting observation here: this implies that the role
> > designated by "undertake" is generic, since in your answer you can
> > pair the noun referent "you" with any other verb.
> Right. However, S11 is *very* young. I made these sample sentences
> just to show that translation isn't really a problem. Whether these
> sentence represent the final structure is not yet decided.
Ahh I see.
> (I must write a Lisp grammar first for experiments, otherwise I
> can't get a good picture of how clause composition feels -- I
> wouldn't want to compose sentences myself, would I? No! Too many
> sandhi rules!)[...]
=-O When I conlang, I do have the aspiration (if not the ability) to
be able to compose sentences on the fly. Tatari Faran has quite a
number of sandhi rules which I sometimes get wrong. But I'm learning.
In fact, I composed the Tatari Faran examples above on the fly. Of
course, there are probably nowhere near as many sandhi rules as your
S11 has, so this may not mean very much. :-P
"I'm running Windows '98." "Yes." "My computer isn't working now." "Yes, you
already said that." -- User-Friendly