Vreleksá Forum Index Vreleksá
The Alurhsa Word for Constructed: Creativity in both scripts and languages
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Logical conlangs
Goto page 1, 2, 3, 4  Next
 
Post new topic   Reply to topic    Vreleksá Forum Index -> Conlangs
View previous topic :: View next topic  
Author Message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Thu Oct 20, 2011 6:21 am    Post subject: Logical conlangs Reply with quote

So I'm still not dead...

Anyway, I've been looking at some interesting papers regarding syntactic verbal structure, and I suppose being in a syntax class helps some Razz
The point is I'm thinking about doing some sort of logical conlang, using a verbal hierarchy/combination structure, maybe approaching oligosynthesis for verbal description.

I'm also wondering if anyone else has worked on (pure-) logic conlangs, and what you did to get around the obvious complications that arise from a different linguistic foundation (ie: a natural non-logical lang).

Any other suggestions?

The paper I've been reading/skimming/not understanding the deep structure of the framework but Really wanting to: "Event Structure and the Encoding of Arguments: The Syntax of the Mandarin and English Verb Phrase" by Jimmy Lin, 2004, MIT (don't know if it's available from a non-subscription source, but hopefully it's accessable).

Thanks in advance!

Back to working on my presentation on Mandarin syntax... :O
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Fri Oct 21, 2011 12:23 am    Post subject: Reply with quote

These don't usually work out.
But the best way to see why is to try it.*
Keep us in the loop, and give it a try.

*(in other words, I, for one, don't know exactly why.)
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Tolkien_Freak



Joined: 26 Jul 2007
Posts: 1231
Location: in front of my computer. always.

PostPosted: Fri Oct 21, 2011 2:37 am    Post subject: Reply with quote

Yeah, I don't know much about logical conlangs at all - it'd actually be cool to see the process of making one, just to see how they work.
Back to top
View user's profile Send private message
Zearen Wover



Joined: 09 Apr 2009
Posts: 42

PostPosted: Fri Oct 21, 2011 3:18 am    Post subject: Reply with quote

I'm not going to pretend I understand half of what you said in that post, but Davin might be considered a logic-language. At the very least, there is a mathematical foundation behind it. It focuses mostly on sets, using owpys (≈verbs) as predicates, and aŋtah (≈nouns) as root sets. As an example, let's study:

Vocab:
(The gloss, placed in brackets will be used as the predicate)
inu [dog]: he is a dog of species ji
eʒalp [tree]: he is a tree of species/kind
efɪs [fast]: he is faster than ji
ɪkej [go]: he goes to ji

let inu efɪs hym eʒalp ɪŋkej.
The dog fast A tree go-TRAN.
The fast dog goes to a tree.

So let and hym are our core sets in this sentence. let is the set of all things in context and hym is the set of all things that exist. It should be noted (though it's not important unless negating), that the first aŋtah in an iprid (noun in a clause) sets the universal set. Next we use owpys to narrow things down. For intransitive owpys, the math basically looks like this:

let inu. => {x ∈ let : dog(x, C)}

Where "C" is the contextual set. So now we have a set of all the dogs in context. Note that this yields another set that we can narrow further with another owpys:

let inu efɪs. => {y ∈ {x ∈ let : dog(x, C)} : fast(y, C)}

So now we restrict it to the set of dogs in context that are also fast. But our definitions can also be transitive, so we have to consider that mathematically. It makes sense that we'd use that set as the second element of the predicate. Thus we finally have:

let inu efɪs hym eʒalp ɪŋkej. => {z ∈ {y ∈ {x ∈ let : dog(x, C)} : fast(y, C)} : ∃ q ∈ {p ∈ hym : tree(p, C)} ∧ go(z, p)}

This is a simplification, some owpys act more like functions that return a modified set, but this holds for most of the common owpys. There was also no discussion of raised phrases which work like lambda expressions. I didn't want to bore you with details.

To address how this helps to solve problems with formal languages vs. natural languages, the language can easily be understood without any of the mathematical background. The only example where the language feels strange do to this foundation is negation.

Also, set specification is a natural place to work from in examining natural language. We use it to set aside parts of the world that we want others to pay attention too. We can even point out things they never realized before. Finally, I kept the syntax as possible, so it doesn't get in the way of communication.

---------

Anyway, I'm really interested in what you're doing, but I don't have the linguistic background to understand your phrasing (yet, but I'm too tired to do research at this time). Also, as someone who has studied Lojban, logic languages definitely interest me. I'll be interested to see what you do Smile
_________________
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Oct 21, 2011 5:39 am    Post subject: Reply with quote

@Zearen: thanks, that's quite interesting! I'm likely to go at it from a logician's point of view as I have a few books I can refer to, and haven't studied formal maths like you mentioned yet.

@eldin: I wonder if it might be that not many conlangers have much practice in formal logic - it would seem to me that there are two types of people who create conlangs: artistic and logical.

The artists try to create an alternate to the mundaneity of natural languages such as English, to be creative, try things, use it in a written setting, etc.

The logicians are fed up with the inconsistencies of natural languages (espeicially English), and try to create an internally consistent method of communication.

The reason I think even the logicians don't turn to pure logic is that unless one has a decent background in it, or an outlined structure in mind, simply using large amounts of cases or agglutination of features like number, gender, etc is more immediately identifiable and satisfying.

@all: I'll post progress when/if I come up with some sort of structure that seems workable.

I'll definitely have another look at Lojban and other loglangs to get a grasp of what's been done. I'll probably take a few ideas from Ithkuil for the morphosyntactic case system, but I think I want to express most of these with the syntactic structure itself.

Should be interesting!
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sat Oct 22, 2011 5:17 pm    Post subject: Reply with quote

Aert wrote:
@eldin: I wonder if it might be that not many conlangers have much practice in formal logic ....

IIRC it's equally because most logicians don't have much practice in conlanging.

If you're a successful conlanger, know something about linguistics, and know formal logic, you may have a chance; certainly a better chance than most, Lojban possibly excepted (though Lojban does have its problems).

Are you planning a taxonomic noun part of your lexicon? That's another thing that hasn't worked out well often.

Are you planning to avoid ambiguity?
One reason that doesn't usually work is that ambiguity is more than a desideratum, it's actually a requirement of any usable language.
Languages won't be usable unless speakers can satisfy Gricean maxims such as "say everything you need to say" and "don't say anything you don't need to say". (Grice put it differently and better, but more formally so it's harder for me to remember.) Ambiguity is built-in to not communicating more information than the speaker needs to tell the listener that the listener doesn't already know and won't figure out for himself/herself.
Also, note, that one reason Lojban fails is that it eliminates as near as possible syntactic ambiguity and morphological ambiguity, whereas IRL most ambiguity is semantic.

If you are careful to formulate your desiderata, preferably as close to "in full" as possible; and record them somewhere so you can remember to check as you go along how close you're coming to them; and maybe also change them (preferably recording the changes) if you change your mind during/because of the work you're doing developing the language; you'll have a better chance.
It might also be a good idea to prioritize your desiderata (that is, your 'lang's design-goals).

It wouldn't hurt our feelings a bit if you posted the design-goals. We can't judge your language ourselves except as it achieves or falls short of your goals.

Good luck, and keep us posted!

Thanks.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Oct 22, 2011 9:25 pm    Post subject: Reply with quote

A taxonomic noun lexicon? Like a hierarchical system? No. I will be using a root system with basic noun types, compounded to get complex nouns. Same with verbs. Right now it looks like adjectives will be derived, adverbs almost definitely will be.

I did actually attempt a hierarchical noun system at one point, and I see your reasoning.

Not sure if I'm going to try to avoid ambiguity completely yet; the conlang will be based on logical formations but I'm not sure if ambiguity will be outright banned = ungrammatical.
I think that things left unspecified would be fine; otherwise the listener would be inundated with irrelevant information every time someone makes a simple statement, not to mention anything complex.

I vaguely recall the maxims you're talking about, and the purpose of this lang, as with all langs, is communication. Only the mode of encoding will be different. Due to the logical base, it may require longer phrases/etc to achieve the same communicative goals, but I see no reason at the moment why it should outright fail.

Thanks for suggesting desiderata though - right now I just have vague working notes.

I suppose my working goals would be as follows:

1. Create a conlang that places heavy emphasis on syntactic structure and relations.

2. This conlang should be close to oligosynthetic in its design, with:
2a. a complex derivational root system for verbs and nouns
2b. derivational structures for other parts of speech
2c. compounding between nouns, nouns and verbs, and possibly other combinations to more specifically express the thought to be communicated
2d. relationship markings on the active phrase constituents.

3. It should be a logically-based language, deriving clauses from an extensive syntactic case-like system, and basic verbal relations to indicate:
3a. constituent relation and coordination
3b. verbal attributes that are otherwise expressed in most natual languages, such as causative, existential attributes, agency, action completion, and others.

4. Develop a complete syntactic description of this conlang, according to Minimalist Theory (which I'm studying), and modify the theory, or use another, such as Event Structure, should Minimalist Theory prove incomplete in its description or structural allowances.

5. Attempt translation from English to this language as proof of concept, then use other natural languages (eg Spanish, Japanese) to compare results of languages I'm not fluent in.

Thanks for the well-wishings, I've got high hopes for this one (as it's not simply another IE-knock-off...)
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Thu Nov 03, 2011 8:05 am    Post subject: Reply with quote

Quick update:

Haven't fleshed out the case system yet, but have made headway in the phonology, and derivational systems.

I'm going to be talking to one of my profs regarding polysynthetic syntactic structures in Athabaskan languages, which I think should help.

I have a few ideas I'm tossing around myself, right now:

1) Derivational root system, using either fundamental verbs and nouns, or category-neutral fundamentals.
Trying to see how to derive basic adjectives; colours are easy 'grass-colour,' etc; some may be able to be derived from the case marking formations, but I need to flesh those out first.
A great deal of compounding will occur, including a common form found in Mandarin known as 'resultative verbal compounding,' in which the active/agentive verb occurs initially, and the resultative verb (what happened) occurs after, eg: 'John hit-break (past) the window' = John hit the window, and it broke = John broke the window (by hitting it).
More details on the roots below:

2) The lang will use a labial/palatal consonant distinction (with no bare form). This could be minimally distinctive for all root words, but it may be used to distinguish the verb-form from the noun-form of the fundamental, non-categorical root (which would never be used in speech). Given current phonotactics (details below). With C-w/j distinction there are almost 6800 syllables of C(C)V(V)(C) structure; without there are just over 3500. Not sure how many fundamental roots will be required for a full lexicon, even one based on oligosynthesis.

3) VSO order, specifically: [tense/aspect]-verb phrase-subject-theme-goal. Theme is the object moved (metaphorical or literal), and goal the destination. Still not sure on the details of even the basic syntactic structure to encode the logicality I'm wanting for the lang.

4) Phonology
Ternary plosive distinction (voiced, voiceless, ejective/extreme aspiration) on labials (b, p, p'); alveolars (d, t, t'), and velars (g, k, k'), as well as uvulars (just q and q').
Fricatives s, z, ʒ, ʃ, x, χ, h
Affricates dz, ts, ts', tɬ, dʒ, tʃ, tʃ'
Approximants/etc l, ɬ, j, w
Nasals m, n, ŋ

Vowels i, e/ɛ, a, o/ɔ, u (x/y contrast is allophonic)
Compounds
Raising ai, ei, oi
Lowering ia, ie, io
Backing iu, au

5) Phonotactics
Basic CV structure, maximally CCVVC per syllable. Permitted clusters listed above.
After uvulars, only a, o/ɔ, u are permitted.
Only finals (codas) permitted are m n ŋ p t k s ʃ l. There is no palatal/labial distinction on finals.
Clusters may only have one supplementary manner of articulation (palatal/labial), such that there is a kind of 'consonant harmony' in clusters: CwC -> CwCw.

I'm looking to find a process to reduce complex (multisyllabic) words (due to the excessive compounding) to di- or trisyllables in order to promote ease of communication. Any suggestions?

6) I will be developing an allophony chart at some point to determine any pronunciation shift in production, but I'm not there yet.


Anyway, I'll get back to you again when I have some better results regarding the syntax or derivational systems...
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Wed Nov 09, 2011 10:54 pm    Post subject: Reply with quote

Well it's 'reading break,' and I've got rather a lot of reading material to work with for the structure of this project Smile

Haven't made any real headway with the root system, but I have figured out the compounding/coordination:

It will definitely use the Mandarin form of resultative compounding, such that v1 is an activity, and v2 the result of the activity. However, this has been expanded to nominal compoundings and secondary results.

The form is now: (v0 topic), v1 activity, v2 result, (v3 result2)
examples:
drop-break: dropped X and it broke
chop-fall: chopped X down

Verbal compounds like "to see X happen," "to want to do" might have to be expressed some other way, be cause v2 is not a result of v1.

Noun-verb compounds: (v0 topic, v1 activity)
picture-paint: paint a picture
paper-write: write a/on paper

As you can guess, these can be combined:

brush-paint-sore [my arm]: paint with a brush (so long that) my arm became sore. In this case, sore would be expressed through an adjectival verb 'hurt.ADJ', and the whole phrase would be expressed in the attributive, which can encode 'become.'

hit-break.ACT-hurt.REFL: broke something by hitting it, and hurt myself.

The reason the noun-verb coordination is backwards compared to the normal verbal expression is to differentiate them in speech:
canvas-paint 1SG.ACT: I canvas-painted (painted a canvas)
paint, 1SG.ACT, canvas.ACC: I painted a canvas
(otherwise the result would be VOS order, which I don't want.)

I was also considering verb-adjective compounds like 'sweep-[CAUS-BECOME]-clean [the floor]' (sweep the floor clean), but I figured that the resultative compounding expressed CAUS-BECOME anyways. So it will be just 'sweep-clean [the floor]' with both sweep and clean as the basic root verbs.


Almost done sorting through the Ithkuil case structures to see what they did to organize everything, mine should be up eventually.


Any suggestions on the syllable-reduction methods? I don't have nearly the experience with phonology that I do syntax...
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Thu Nov 10, 2011 7:44 pm    Post subject: Reply with quote

Aert wrote:
With C-w/j distinction there are almost 6800 syllables of C(C)V(V)(C) structure; without there are just over 3500. Not sure how many fundamental roots will be required for a full lexicon, even one based on oligosynthesis.
3500 is probably enough. 6800 is almost certainly enough.


Aert wrote:
Theme is the object moved (metaphorical or literal)

"Theme" is the entity moved ("Gipper threw the ball") or located ("Here is the ball"). In those two examples, literally. By "metaphorically" I'm guessing you're going by the "lexicalist hypothesis", which if I understand correctly is that states-of-being "stand in for" places or locations; so for instance "Gipper gave the ball to Knute") has "the ball" being moved to the destination "Knute".


Aert wrote:
I'm looking to find a process to reduce complex (multisyllabic) words (due to the excessive compounding) to di- or trisyllables in order to promote ease of communication. Any suggestions?

(1) Reduce all consonants of any particular MoA (Manner-of-Articulation) in non-stressed syllable's onset to the same PoA (Place-of-Articulation).
AND/OR;
(2) Reduce all consonants of any particular MoA (Manner-of-Articulation) in non-stressed syllable's coda to the same PoA (Place-of-Articulation).
AND/OR;
(3) Reduce any vowel cluster in any non-stressed syllable to a single vowel.
AND/OR;
(4) In any word of four or more syllables in which there are two adjacent non-stressed syllables, leave the vowel(s) out of one of them.
AND/OR;
(5) In any word with a cluster of four or more consonants, leave out all but three of them.

Truth is, I don't know. But those are some ideas. Don't know how good they are.

Maybe you'll find some hints somewhere in this pdf.

Or this.

Or http://www.sciecom.org/index.php/LWPL/article/download/2359/1934.

Or a few other links that pop up in this search.

these guys wrote:
unstressed syllables are omitted more often in weak-strong (iambic) than in strong-weak (trochaic) environments

_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Feb 25, 2012 8:42 am    Post subject: Reply with quote

Hi again,

I've been thinking about this conlang again, especially after a few specific discussions in my syntax class.

We were discussing A-not-A phrases (common in Mandarin) which is more or less like English "Do you like X or not?", expressed "you like-not-like X?". Given that Mandarin is SVO, like English, I wondered what it would look like in a VSO, like this one.

I also recalled we were talking about AgrO (verb agreement with object) nodes in syntactic trees, so I threw that in there.

Then I remembered Mandarin enjoys using serial verb constructions, especially resultative "chop-fall" = chopped (at) the tree and (as a result) it fell. This was already on the board for my conlang, so I looked at how all these things can be combined into a single syntactic phrase structure.

The basic trees are here:
http://www.flickr.com/photos/28228492@N05/6781200756/in/photostream
SVO and VSO orders are marked, with the more complex structures given in VSO order only.

With AgrO in SVCs, I've decided that the following structures will be allowed:
- all intransitive verbs
- monotransitive verbs
- ditransitive verbs only when the first object has different phi-features (gender, number, person, etc) than the second; or when the two actions are being done to the same object. Too much ambiguity.

eg:
intransitive: eat-finish [...]
monotransitive: wash_it-clean [...]; cook_it-eat
ditransitive (acceptable): take_itA-throw_itA [...] "took the ball and threw it"
ditransitive (acceptable): invite_them-eat_it [...] "invited them to eat dinner"
ditransitive (unacceptable): take_itA-cut_itB [...] "took a knife and cut the bread"

In any case, there are other ways of expressing unacceptable ditransitive SVC phrases, eg. via subordination, or using case markers on nouns (eg. "cut-it [Subj] knife.INSTR bread.ACC").


I might also to reduce the logical verbal constructions (eg. "CAUSE-MOVE-[verb] [subject] [object]" as it's rather redundant, but also so I can use the important (distinctive) verbal modifiers as suffixes on the main verb. I'm not entirely sure on this one, because I can't remember if a VSO language is more likely to have prefixes or suffixes.


There are probably other things rattling around in there, but I have midterms to study for.
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sat Feb 25, 2012 9:32 pm    Post subject: Reply with quote

I am going to have to think and/or look up some terms to understand what you just said.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Tue Feb 28, 2012 12:31 am    Post subject: Reply with quote

Thought it over - why would there be object agreement without subject agreement?

So with subject agreement, overt pronunciation of the subject itself would be optional (cf. in Spanish: "(yo) hablo espanol").

I also noticed that in a logical VSO language, negation should precede the phrase it negates.

Looking at the AgrO1,2 nodes on the tree, it makes more sense to have them on the separate verbs in a serial verb construction, otherwise (with longer combinations), it would get very messy. Also, this way is nice and symmetrical, and rather elegant, I think Smile

Since subject precedes object in VSO, and taking into account the repositioning of negation, the new syntactic tree is as follows:

CP > PolP > TP > AgrS > AgrO1 > vP > AgrO2 > VP (> CP for embedded clauses).

The full structure:

(INT) (NEG) TENSE v1 AgrS AgrO1 AGENT v2 AgrO2 OBJECT (GOAL)

Example:

INT NEG PAST chop-you-(it) you fall-(it) tree
"(Is it true that) you did not chop the tree down?"

In this way, serial verb constructions are easy to coordinate with the proper agreement nodes, and standard (non-serial) verb constructions behave exactly the same, except a second verb is not left in AgrO2.


With regard to A-not-A structures, that has to be rethought given the repositioning of the negation node.

As a result, I've simply motivated that the verb, once in place at AgrS, copies itself to T before the TENSE marker. This gives [(INT) NEG v1 TENSE v2 ...]. I'm not yet sure how conjunction phrases, eg "do you like or love her?" will work, as I haven't studied conjunctions in detail yet.


Given that the standard structure is the serial-verb construction without the second verb, I'll post the image for the SVC for reference:
http://www.flickr.com/photos/28228492@N05/6936706545/in/photostream
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Thu Mar 01, 2012 12:39 am    Post subject: Reply with quote

Did a few more edits:

1) Moved PolP back below TP - forcing a particular reading of a negation phrase, while also allowing A-not-A phrases.

2) Expanded PolP to include the emphatic marker (eg "I DID go to the library").

3) Reworked where secondary verbs are generated, which makes for a more elegant movement pattern.

4) Formalized the various structures into non-tree descriptions:

Standard
V1-AgrS (AgrO) AGENT OBJECT <V1> (GOAL)
VSO

Serial-verb construction
V1-AgrS (AgrO1) AGENT V2 (AgrO2) <V2> OBJECT <V1> (GOAL)
V1SV2O

Noun Incorporation*
V1 (AgrS) OBJECT AGENT <Obj> <V1> (GOAL)
VOS

A-not-A**
V1 NEG _V1_-AgrS (AgrO) AGENT OBJECT <V1> (GOAL)
V1-not-V1 SO (VSO)

Questions: wh-in-situ (not moved)
QP V1-AgrS (AgrO) AGENT OBJECT=WH <V1> (GOAL)
VS WH (VSO)

Questions: wh-movment
OBJECT=WH QP V1-AgrS (AgrO) AGENT <v1> <Obj> (GOAL)
WH VS (OVS)

Notes on how these examples are written:
anything inside () is optional.
anything inside <> has been moved up (left)
anything inside _ _ indicates copied movement up (left) [original item stays there, with a copy of it in the moved position]

*note that in noun incorporation, the thematic roles (arguments the verb can take) is reduced by one. This is done here by moving the object into the position of AgrO1 (which normally only agrees with the object).
As a result, it is ungrammatical to use AgrS in incorporation structures, because it would get in the way of the verb and incorporated object noun. (This is marked by underlining, as strikethrough isn't an option on this forum.)

** note that in A-not-A, the verb is copied from its position in AgrS, not moved: it is generated down in V, and moves up to AgrS, and then (in A-not-A only) copied to T.

Legend:
V1 main verb
V2 second verb (in SVC, including resultatives, etc)
AgrS agreement marking with the subject/agent of the phrase
AgrO1 agreement marking with the object of the phrase
AgrO2 agreement marking with the secondary object of the phrase (uncommon)
QP question particle, like Japanese 'ka'
WH a wh-phrase (question word/interrogative marker)

---

I'm also thinking about using alternative marking, or 'disjunct' questions, eg: 'X or Y ...': "Do you like her or love her?" Not sure how to go about this just yet though, unless I can use PolP to house a conjunction...

This is starting to look good Smile
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Fri Mar 02, 2012 8:59 pm    Post subject: Reply with quote

Aert wrote:
Thought it over - why would there be object agreement without subject agreement?

For a finite verb that's actually the nucleus of a clause, I'd guess there wouldn't be, unless one of the "agreement with the subject" morphemes happens to be a "zero-morpheme".

But for a passive participle, generally if adjectives agree with their head-nouns, the passive participle will agree with its "object" (patient) but not its agent (which probably doesn't even show up explicitly in the utterance); even if finite verbs agree only with their subjects.


Aert wrote:
This is starting to look good Smile

Yes. Cool
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Mar 09, 2012 11:57 pm    Post subject: Reply with quote

Did some fine-tuning on the serial verb constructions, and the (seemingly final) formula for sentence-level (CP) expression is:

CP > TP > [V(N) (Agr(Args)) Arg(Args)]N
Args = (S(O(G)))

This doesn't show up very well on the forum, so here's the explanation:
We have a verb, which is numbered (N).
We have optional agreement structures (I'll expand more on this later).
We have arguments (minimum one: the subject).
This level can be iterated (the N at the end, which corresponds to the numbered verb).

The arguments are: subject, object, goal (roughly, NOM, ACC, DAT cases). Note the embedded structure, such that you would not be able to have an object without first mentioning the subject, or a goal without the object (which would also require the subject).

Since this language is VSO, each verb-level phrase comes out: "verb-(agreements) S O (O2, etc)"


I noticed that there are really two different SVCs, which are in complementary distribution regarding grammaticality. We then have:

Resultative (chop down a tree: "chop I fall tree" = V S V O)
[V1 (Agree) Subj *Obj *Goal]1
[V2 Agree *Subj Obj (Goal)]2
with any further serial verbs following the structure of (2).

Note that in the first verbal phrase, the object and (optional) goal cannot occur. In contrast, the second verbal phrase requires verbal agreement with the subject and all other arguments, (keeping a VSO structure at the verbal agreement level), but disallows the subject proper, because 1) the subject is already indicated in the agreement, and 2) this is the object verbal phrase of a resultative clause, indicating that the action of the verb is happening/experienced by the object.

Example:

[chop-(1sg-(3sg.n)) 1SG] [fall-1sg-3sg.n-(3sg.pl) tree firewood.DAT]
"I chopped the tree down for firewood" ["I chopped (at) the tree, and as a result, it fell; I will use it for firewood"]

Coordinative (borrow a pen to write a letter: "borrow pen write letter" = V S V O)
[V1 (Agree) Subj Obj (Goal)]1
[V2 Agree *Subj Obj (Goal)]2
again, any further serial verbs would follow the structure in (2).

Note that in coordinative SVCs, objects and goals are allowed in the primary verb phrase. The structure for the secondary verb phrase(s) is the same in both coordinative and resultative SVCs.

Example:
[borrow-(1sg)-(3sg.n) 1SG pen] [write-1sg-3sg letter]

another:
[catch-(1sg)-(3sg.n) 1SG bus] [watch-1sg-3sg.n movie] [eat-1sg-3sg.n dinner] [fall_asleep-1sg-3sg.n]
"I caught a bus, watched a movie, ate dinner, and fell asleep"
(literal coordination).


There are two options for SVC structures: recursive subject agreement, and full agreement. This is the reason for optional subject agreement in the first verb phrase.

Recursive subject agreement says that every verb takes the agreement marker related to the subject. And, since there is subject agreement (but no object/goal agreement for resultative SVCs), the subject may be dropped (cf. Spanish: (yo) hablo).

Full agreement is just that: all arguments of each verb phrase illicit agreement on the relevant verb.

Quote:
This is starting to look good.
Quote:
Yes Cool

Thanks Smile


Think that's all for now - I need food...
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Mar 10, 2012 6:37 am    Post subject: Reply with quote

Edit: I just found out that serial verb constructions can Only occur in languages where the tense is not realized on the verb. Since this language has the tense expressed at the head of TP, and Not on the verb, it is perfectly within the narrow definition that these structures can happen!

*whew!
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Tue Mar 13, 2012 6:48 pm    Post subject: Reply with quote

I think I'm going to have to go with a theory stating that a single clause phrase can hold no more than one matrix verb (modals and auxiliaries like 'do,' 'be,' etc aren't counted). This way there is no complex syntactic structure as to why there are serial verbs, but the trees will be huge.

As a complement to the complexity of verb constructions in this lang, I'm distinguishing between syntactic SVCs and morphological verbal compounds. Verbal compounds contain one or more verbal affixes, plus the matrix verb root. The affixes encode things like locative and motion-related markings, and modal verbs including 'want,' the copula, and etc.

Not sure yet if the morphological constructions will require agreement of form on serial verbs or not. If they encode tense/aspect indicators, it's possible that they could be copied. Other affixes would be extremely unlikely to undergo copy movement.

Unfortunately, there's a theory that says that syntactic operations (eg copy movement) can't look inside a word item (eg complex verb). If this holds true and I intend on having tense/aspect agreement from morphological affixes, they would have to really be generated on separate syntactic nodes, and thus not be morphological in nature. They would likely even move up to TP (where the tense aspect marker (TAM) is located). This would result in a complex TAM, which would be the most likely item for copy agreement across clauses in SVCs.

Hope you don`t mind me using this board as a notepad for what comes out of my scribblings; I have yet to make a proper grammar document, and it`s nice to get feedback Smile

Cheers!

EDIT: I`m also rethinking my position on using A-not-A. I`m doing a paper on the topic, and so far the only languages I`ve seen it in are underlyingly SOV. I`ll have to do a bit more digging as to opinions on how it`s generated, but it may still be an option.
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Mon Mar 19, 2012 4:09 am    Post subject: Reply with quote

Damn - apparently serial verb constructions are specifically NOT across clauses (eg. embedded). Some proposals are saying they're made with conjunctions which aren't pronounced, but that would only (possibly) work with the coordinative SVCs, not resultatives.

I'm going to have to do some more reading about how complex SVCs could be made - I don't want to have to limit the recursive nature of these, but a non-clausal syntactic structure for these would get really complicated. At the same time I know the CP/TP levels would get in the way unless I specified that the tense marker moves from where it's generated up to the highest TP and pronounced there; question words and etc would similarly move (as they do under normal assumptions).

Oh well...
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Mon Mar 19, 2012 9:52 pm    Post subject: Reply with quote

I'd say the unspoken conjunctions in SVCs are either "then" or "and".
(Actually I'd say the idea of unspoken conjunctions being necessary or even allowable is a typical Generativist tic.)

Could you have been thinking of clause-chaining with switch-reference marking, rather than of serial-verb constructions?

When clauses are chained together in clause-chaining languages, it's possible that one is "embedded" in the other, but as the first or the last element.

All the clauses could be considered dependent on the anchor clause of the chain, which could therefore be considered the main clause.

Chaining clauses together is something that's fuzzy between co-ordinating conjunction and sub-ordination.

Since clauses are usually chained in "iconic" order, their tense is usually dependent on the anchor clause's tense.

A switch-reference system could also make the participants of one clause depend on the participants of one of its neighbors or of the anchor clause.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    Vreleksá Forum Index -> Conlangs
Goto page 1, 2, 3, 4  Next
All times are GMT
Goto page 1, 2, 3, 4  Next
Page 1 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2002 phpBB Group
Theme ACID © 2003 par HEDONISM Web Hosting Directory


Start Your Own Video Sharing Site

Free Web Hosting | Free Forum Hosting | FlashWebHost.com | Image Hosting | Photo Gallery | FreeMarriage.com

Powered by PhpBBweb.com, setup your forum now!
For Support, visit Forums.BizHat.com