Page 2 of 2

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Sun Jan 16, 2011 11:21 pm
by vec
Ah, I see. The thing is, when reading a document like this on a screen, the footnotes get a little lost because you can't refer to them without scrolling. Maybe you should consider moving some of them into the text? Or even use sidenotes? Those are harder to manage, but easier to read. Maybe the cross-references are moved into the text, the other notes kept as footnotes. I don't know. Or I can just learn to live with some scrolling.
I'm going to read some more now.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Mon Jan 17, 2011 6:11 am
by Jipí
Well I hope I feel like writing chapter 1 and the ones after 6 sometime this year …

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Mon Jan 17, 2011 4:01 pm
by vec
So, finished reading through. Nothing of note that I see. I can't wait for more!

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Mon Jan 17, 2011 4:08 pm
by Jipí
Great job, thanks!

Don't you think I could elaborate a bit more on basic morphosyntactic alignment chapter? I mean, what I wrote down there already is all I could think of that was relevant for that chapter, as Descr. Morph. has it. But yeah, anything of interst to add there?

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Mon Jan 17, 2011 7:04 pm
by vec
Yes, definitely. I assumed it wasn't complete. I also wouldn't mind some declension tables as an overview to better visualise the relationship between different marked categories. I'm honestly not sure what chapter they should go in, though. I've never read Describing Morphosyntax, so the order in which you describe things is a bit alien to me, as I said before. But as I got into it was fine. But I do feel like the grammar is missing some sort of overview chapter of declension patterns.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Mon Jan 17, 2011 7:11 pm
by TomHChappell
I'll be interested in chapters 9 through 12.

I think you should be congratulated for doing chapters 1 through 8 as thoroughly as you have.
Is the kind of focus marker on the verb Ayeri has, common in natlangs? Which natlangs have such a thing?

In a causative clause, is the agent-of-effect marked as if it's an instrument?

It's interesting that genitive=ablative (sources or origins). Is that as common in natlangs as, or more common than, dative=allative (goals or destinations, 64b in 5.4)? (BTW I do not see 64a of 5.4 as a good example of "locational genitive". "I tell of my work" doesn't seem very ablative. A genitive construction for "The girl from Ipanema" would be better IMO.)

One could and might expect the thing located in 7.2 predicate locatives to be marked the same as the thing transferred in ditransitive sentences. In both cases it has the semantic role of "theme" (thing located or moved). But is that what happens? It looks like you've got it marked as Agent in 7.2; if Ayeri's case-marking is more semantic role than grammatical relation, the thing located is hardly an Agent, is it?

Have you explained your glossing abbreviations yet? You probably need another appendix to do that. Right now it could probably be a lot shorter than your other appendices. Right now the fact that A is Agent and P is Patient is noted in the first time you use them, but not in a "central" location that collects all of the glossing abbreviations.

You've spelled "denominalization" two different ways (also "denominalisation"); American with <z> and British with <s>. And you haven't defined it. "Nominalization" is making a noun out of something. I thought a "denominal verb" was a verb derived from a noun but retaining no noun-like features; a "denominal adjective" was an adjective derived from a noun but retaining no nounlike features; and a "denominal adverb" was an adverb derived from a noun but retaining no nounlike features; and so on in case there is such a thing as a "denominal adposition" or a "denominal conjunction". If that's not denominalization, what is? http://www.cis.upenn.edu/~sriramv/emnlp ... _joshi.ppt mentions "give a smile", "take a look", "pay a visit". Are those the denominalizations you were talking about?

Why are those mentioned in the same breath as nominalization? In what way are the processes parallel or opposed, rather than their merely being a similarity in spelling and pronunciation of the terms by which those processes are called?

In 6.1 about compounds, can you derive each some word of each part-of-speech from a compound of two words of each part-of-speech? If you have seven parts of speech (noun, verb, adjective, adverb, adposition, conjunction, pronoun) are all combinations of part1 <-- part2 + part3 possible?
Just looking at noun <-- noun+noun for the moment; Panini mentioned six types, if I recall correctly. Among them were dvandva, tatsuma, and bahuvrihi. Are you going to cover all six (if that's the right number) types in 6.1? Do all types occur in Ayeri or are there some of Panini's types that just don't apply to Ayeri?
How about
verb <-- verb+noun
or
verb <-- verb+adposition
or
verb <-- verb+verb
compounds?
I know 6.1 is part of 6 "Noun and Nounphrase Operations", so
word <-- noun+noun
and
noun <-- word+word
compounds are the kind 6.1 is (probably entirely, but at least) mostly about. Nevertheless, ...

Can you express all 30 of Fillmore's "The Case for Case" roles in Ayeri via some combination of Ayeri cases, prepositions, postpositions, and focus-markers? Does every combination of a case and a preposition have a meaning? Will there be a table somewhere that will say what those are? Do some of them also depend on the focus-marker for their specific meaning? Can you use a preposition and a postposition both at the same time on the same object noun-phrase as long as it's in locative case?

4.4.2 how about fractions? "half, one-and-a-half, one-third, one-less-a-third, one-and-a-third, quarter, one-less-a-quarter, one-and-a-quarter, etc.". You have multiplicative numbers like twice and thrice etc.; can you also have numbers that multiply by a fraction?

I'm running out of time. I still have to finish reviewing chapters 1-5. But that's what I've got to ask about chapters 6-8 so far. Sorry I haven't proof-read my questions; some of them may not be well-asked, for which I apologize if that's so.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Tue Jan 18, 2011 3:38 am
by Jipí
vecfaranti wrote:I also wouldn't mind some declension tables as an overview to better visualise the relationship between different marked categories.
Well, there is an overview at the beginning of the chapter on nouns and on verbs of all the slots inflection can get into. Or do you mean a one-page quick reference? Because there isn't much to inflect, basically just nouns (case, number), verbs (tense, mood, progressive aspect, person(/gender/number/case)), pronouns (gender/class, number, case), and relative pronouns (case of the referent NP, case of the relative pronoun). And those all have just one paradigm each.
TomHChappell wrote:Is the kind of focus marker on the verb Ayeri has, common in natlangs? Which natlangs have such a thing?
I think the letter question would be a case for either the introductory chapter, or the chapter that will deal with this.
In a causative clause, is the agent-of-effect marked as if it's an instrument?
Care to explain agent-of-effect?
It's interesting that genitive=ablative (sources or origins). Is that as common in natlangs as, or more common than, dative=allative (goals or destinations, 64b in 5.4)? (BTW I do not see 64a of 5.4 as a good example of "locational genitive". "I tell of my work" doesn't seem very ablative. A genitive construction for "The girl from Ipanema" would be better IMO.)
Hm, I've not done comparative research. Also ,yes, "from Ipanema" would be more prototypical than "of work".
One could and might expect the thing located in 7.2 predicate locatives to be marked the same as the thing transferred in ditransitive sentences. In both cases it has the semantic role of "theme" (thing located or moved). But is that what happens? It looks like you've got it marked as Agent in 7.2; if Ayeri's case-marking is more semantic role than grammatical relation, the thing located is hardly an Agent, is it?
If you mean the "The book is for him" example, I marked that with an asterisk anyway. I'd still, however, mark "the book" as an agent, but I can see how that is basically the same thing as the theme in ditransitive clauses. On the other hand, why can't it be a little illogical in places? This is in analogy with other predicative clauses rather than with ditransitive clauses.
Have you explained your glossing abbreviations yet? You probably need another appendix to do that.
Yes. It's right after the overview of tables and illustrations, on page 5 of the file (the page in print is unnumbered).
You've spelled "denominalization" two different ways (also "denominalisation"); American with <z> and British with <s>. And you haven't defined it.
My bad.
"Nominalization" is making a noun out of something. I thought a "denominal verb" was a verb derived from a noun but retaining no noun-like features; a "denominal adjective" was an adjective derived from a noun but retaining no nounlike features; and a "denominal adverb" was an adverb derived from a noun but retaining no nounlike features; and so on in case there is such a thing as a "denominal adposition" or a "denominal conjunction". If that's not denominalization, what is? http://www.cis.upenn.edu/~sriramv/emnlp ... _joshi.ppt mentions "give a smile", "take a look", "pay a visit". Are those the denominalizations you were talking about? Why are those mentioned in the same breath as nominalization? In what way are the processes parallel or opposed, rather than their merely being a similarity in spelling and pronunciation of the terms by which those processes are called?
Got to have a look at that again. Thanks for the notice.

Gotta go to uni now. I'll have a look at the other points you mentioned later.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Tue Jan 18, 2011 3:21 pm
by Jipí
Guitarplayer wrote:Got to have a look at that again. Thanks for the notice.
I am used to refer to any process that derives a noun from another part of speech as "nominalization". Whether that's deverbal, deadjectival, or de-whateveral. Those I'd handle as subclasses of nominalization if it's worth to point out differences.
TomHChappell wrote:In 6.1 about compounds, can you derive each some word of each part-of-speech from a compound of two words of each part-of-speech?
TBH, I didn't find examples for all 49 combinations of word+word. Like, marinsoyang, lit. 'in-front-or'. Great, what use would that be?! Do I explicitly have to mention which combinations don't appear? Also, this is a conlang, so its lexicon is still growing, so I don't know what I'll be doing in 1..5..10 years, should I still be working actively working on this then. Of course, this doesn't mean I can't go back and revise things. As for Pāṇini's types of compounds, I made up at least one dvandva (tihandekey, 'cutlery' < tihang 'knife' + dekey 'fork', however for some reason this hasn't ended up in the dictionary, or got lost by transferring between an online and an offline copy of my dictionary database both ways) and one or two bahuvrihis (that I don't remember anymore). Strictly speaking, though, the dvandva is not a pure one, as that would require both words to be in the dual (AFAIK), which my conlang doesn't have.
Can you express all 30 of Fillmore's "The Case for Case" roles in Ayeri via some combination of Ayeri cases, prepositions, postpositions, and focus-markers? Does every combination of a case and a preposition have a meaning? Will there be a table somewhere that will say what those are? Do some of them also depend on the focus-marker for their specific meaning? Can you use a preposition and a postposition both at the same time on the same object noun-phrase as long as it's in locative case?
This requires further investigation. However, I, too, think that too much "cartesianness" in conlangs (as exemplified and criticized by Miekko) is a pitfall that should be avoided. And that I'm probably not unguilty of.
4.4.2 how about fractions? "half, one-and-a-half, one-third, one-less-a-third, one-and-a-third, quarter, one-less-a-quarter, one-and-a-quarter, etc.". You have multiplicative numbers like twice and thrice etc.; can you also have numbers that multiply by a fraction?
I don't see how the last question is practical (how often do you say "three-quarter-times"?). It would be fun to figure out, though. As for the other question about integer ± fraction, I sketched out mathematical terminology before. I think it would indeed be worthy to add what I've worked out so far.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Wed Jan 19, 2011 8:21 am
by Jipí
A problem that's been bugging me more than compounding possibilities is how to express more complex causatives. In chapter 6.4.7 there are three examples that are important in the context I wonder about:
(a) Ang rua sara-ayn-Ø seyaran-isa.
AF must leave-1p.FOC rain-CAU
‘We had to leave due to rain.’

(b) Sā no sip-ara adanya-Ø vās.
CAUF want cry-3si that-FOC 2s.P
‘It makes you want to cry.’

(c) Sā apa-yang petān-Ø yana.
CAUF laugh-1s.A folly-FOC 3sm.GEN
‘I laugh due to his folly.’
Now, the point is, (b) explains how to do A makes B modal verb, however what happens if verb is transitive? Which case should that third constituent NP have? Dative, Genitive, Instrumentative, Locative? I think this depends heavily on the semantics of the verb. Also, does it really make sense to have the one forced to act (thus in fact an Agent, even though an indirect one!) be the Patient? Also, the verb usually agrees with the Agent, why does it not here?

A solution might be to go by analogy of (a): AF MOD verb-AGR\B B.FOC A.CAUCAUF MOD verb-AGR\B B.AGT A.FOC (which is the structure in (c), basically!), so what had been oblique before is now demoted to focus, hence forming kind of an applicative, which would be within the principles of the lang. For three constituents we'd thus get CAUF MOD verb-AGR\B B.AGT C.* A.FOC 'It is A who causes B to modal verb C', or 'A makes B modal verb C'. Comments?

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Sun Jan 23, 2011 3:19 pm
by Jipí
Guitarplayer wrote:A solution might be to go by analogy of (a): AF MOD verb-AGR\B B.FOC A.CAUCAUF MOD verb-AGR\B B.AGT A.FOC (which is the structure in (c), basically!), so what had been oblique before is now demoted to focus, hence forming kind of an applicative, which would be within the principles of the lang. For three constituents we'd thus get CAUF MOD verb-AGR\B B.AGT C.* A.FOC 'It is A who causes B to modal verb C', or 'A makes B modal verb C'. Comments?
I simply did this now because it seemed more logical to me than the way I had had it before.

Re: Ayeri grammar thread - updated 01/14/2011

Posted: Thu Jan 27, 2011 3:37 pm
by TomHChappell
Guitarplayer wrote:
In a causative clause, is the agent-of-effect marked as if it's an instrument?
Care to explain agent-of-effect?
Agents control and/or perform and/or effect and/or instigate the action the clause describes.
In causative clauses the "agent" role is split; there are typically two agents, an instigator and a performer.
If A makes B do C to D, A is the initial instigator and B is the final performer. Both of them may exercise more or less control and may more or less effect the action.
The "initial instigator" or "causer" or "agent-of-cause" is A.
The "final performer" or "causee" or "agent-of-effect" is B.
Guitarplayer wrote:Hm, I've not done comparative research
I don't mean to embarass you; I haven't done any research either. I'd just be curious to know the answer.
Guitarplayer wrote:If you mean the "The book is for him" example, I marked that with an asterisk anyway. I'd still, however, mark "the book" as an agent, but I can see how that is basically the same thing as the theme in ditransitive clauses. On the other hand, why can't it be a little illogical in places? This is in analogy with other predicative clauses rather than with ditransitive clauses.
A perfectly acceptable explanation.
Guitarplayer wrote:
Have you explained your glossing abbreviations yet?
Yes. It's right after the overview of tables and illustrations, on page 5 of the file (the page in print is unnumbered).
Oops! :oops: I missed it. Yes, it's there, and good.
Guitarplayer wrote:I am used to refer to any process that derives a noun from another part of speech as "nominalization". Whether that's deverbal, deadjectival, or de-whateveral. Those I'd handle as subclasses of nominalization if it's worth to point out differences.
That makes sense. My question was, "what did you mean by denominalization?". Let us know your thoughts when you want to.
Guitarplayer wrote:
TomHChappell wrote:In 6.1 about compounds, can you derive each some word of each part-of-speech from a compound of two words of each part-of-speech?
TBH, I didn't find examples for all 49 combinations of word+word. Like, marinsoyang, lit. 'in-front-or'. Great, what use would that be?! Do I explicitly have to mention which combinations don't appear? Also, this is a conlang, so its lexicon is still growing, so I don't know what I'll be doing in 1..5..10 years, should I still be working actively working on this then. Of course, this doesn't mean I can't go back and revise things.
No, I don't think you should explicitly mention which ones don't occur. But I would like you to mention (and provide examples) of those that do occur.
OK, let me do some thinking and just type my thoughts here; I hope it'll be a bit better organized than "stream-of-consciousness", but I apologize if it isn't.
When I was in school most Americans were taught that English has eight word-classes (parts-of-speech); to wit and namely, in (nearly) alphabetical order,
  • adjectives,
  • adpositions,
  • adverbs,
  • conjunctions,
  • nouns,
  • pronouns,
  • verbs,
  • and interjections.
I'll leave interjections out of consideration altogether.
There are then 21 possible types of source-pairs (disregarding order);
  • adjective+adjective
  • adjective+adposition
  • adjective+adverb
  • adjective+conjunction
  • adjective+noun
  • adjective+pronoun
  • adjective+verb
  • adposition+adposition
  • adposition+adverb
  • adposition+conjunction
  • adposition+noun
  • adposition+pronoun
  • adposition+verb
  • adverb+adverb
  • adverb+conjunction
  • adverb+noun
  • adverb+pronoun
  • adverb+verb
  • conjunction+conjunction
  • conjunction+noun
  • conjunction+pronoun
  • conjunction+verb
  • noun+noun
  • noun+pronoun
  • noun+verb
  • pronoun+pronoun
  • pronoun+verb
  • verb+verb
I think "into" and "onto" could be English adpositions that are formed by combining two adpositions.
What part-of-speech are Spanish's "conmigo" and "contigo"? Could they be considered adposition+pronoun compounds?
English has lots of verb+adposition "two-part" verbs, pretty much equivalent to German's "separable" verbs. But "income" is a noun that's an adposition+verb combination.
Isn't it possible and reasonable and plausible to consider English's "and/or" to be a conjunction that's a compound of two conjunctions?
English occasionally has objects or patients incorporated into verbs; it also sometimes has subjects or agents incorporated into verbs.
There's a pidgin of English that has the inclusive 1st+2nd person dual pronoun "yumi" and the 1st+2nd person inclusive trial pronoun "yumitripela". I think "yumi" is a compound of "yu" and "mi".
Some people might say "gimme" and "I'd've" and so on are verb+pronoun or pronoun+verb compounds in English; but maybe not.
Possibly not all of those pairs would reasonably occur in a given language.
Possibly some could occur but the resulting compound could only be one part-of-speech; that's likely to be true of, for instance:
adposition+adposition --> adposition
conjunction+conjunction --> conjunction
pronoun+pronoun --> pronoun
Endocentric compounds, dvandvas, and appositional compounds, are likely to be the same part-of-speech as the head constituent. But exocentric compounds such as bahuvrihis might not be.
But there's a French verb, "tutoyer", derived from a pronoun+pronoun compound. And English has a noun, "voirdire", derived from a French verb+verb compound. So I'm not certain I can say for sure that if two words of a given class are compounded, the resulting compound word must be of the same class.
http://en.wikipedia.org/wiki/Compound_% ... _compounds mentions nouns like "corkscrew" that are compounded of a verb and a noun.
Guitarplayer wrote:As for Pāṇini's types of compounds, I made up at least one dvandva (tihandekey, 'cutlery' < tihang 'knife' + dekey 'fork', however for some reason this hasn't ended up in the dictionary, or got lost by transferring between an online and an offline copy of my dictionary database both ways) and one or two bahuvrihis (that I don't remember anymore).
I'd like to see them if you can find or remember or re-create them.
Guitarplayer wrote:Strictly speaking, though, the dvandva is not a pure one, as that would require both words to be in the dual (AFAIK), which my conlang doesn't have.
I thought a dvandva was just any two-headed compound noun; that is, you combine two nouns, and either one of them is just as much the head of the compound as the other. I never heard that it had anything to do with the dual grammatical number.
See http://en.wikipedia.org/wiki/Dvandva. It looks like the dvandva ordinarily lacks a singular (except for samāhāra-type dvandva which are collective nouns), but in Sanskrit it can have both a dual and a plural. The English example given, "singer-songwriter", is not a samāhāra because it isn't collective, and is not an ekaśeşa because neither source-noun is mostly omitted, but it's not an itaretara either because the singer and the songwriter are the same person so it's singular instead of dual or plural. I'd say "singer-songwriter" is a dvandva to the extent that it's a compound at all; both "singer" and "songwriter" are heads together of this compound, (if it's a compound), and that's what makes it a dvandva. However, http://en.wikipedia.org/wiki/Compound_% ... sification calls "actor-director" and "maidservant" appositional compounds rather than dvandvas. "Bittersweet" (adj+adj=adj) and "sleepwalk" (verb+verb=verb) are probably better English dvandvas.

What about tatpuruṣa or endocentric compounds? "Housewife", "doghouse", "blackbird", etc.?
Guitarplayer wrote:
Does every combination of a case and a preposition have a meaning? Will there be a table somewhere that will say what those are? Do some of them also depend on the focus-marker for their specific meaning? Can you use a preposition and a postposition both at the same time on the same object noun-phrase as long as it's in locative case?
This requires further investigation. However, I, too, think that too much "cartesianness" in conlangs (as exemplified and criticized by Miekko) is a pitfall that should be avoided. And that I'm probably not unguilty of.
OK.
Guitarplayer wrote:
me wrote:Can you express all 30 of Fillmore's "The Case for Case" roles in Ayeri via some combination of Ayeri cases, prepositions, postpositions, and focus-markers?
However, this part of my question is not really about "Cartesianness"; it's about expressive ability. I'm sure that if your conlang is complete, or once it is complete, you can expressd all 30 of Fillmore's whatever-those-things-were (thematic relations, perhaps?); my question is, do you intend to be able to do them all by combining cases, prepositions, postpositions, and focus-markers? And if so, can you do them all yet?
I wasn't really interested in whether or not every such combination has a meaning ("cartesianness"); but, rather, in how you expressed each of those meanings, whether via an appropriate combination or otherwise.
Guitarplayer wrote:
4.4.2 how about fractions? "half, one-and-a-half, one-third, one-less-a-third, one-and-a-third, quarter, one-less-a-quarter, one-and-a-quarter, etc.".
As for the other question about integer ± fraction, I sketched out mathematical terminology before. I think it would indeed be worthy to add what I've worked out so far.
I guess I missed your mathematical terminology. If I don't find it by looking again, can you point me to it?
Latin has all kinds of "sesqui-" numerical prefixes. "Sesqui-" comes from "semis+que" where "que" is the enclitic form of Latin's conjunction meaning "and". "Sesqui-" by itself means one-and-a-half; "sequidi-" means two-and-a-half; "sesquitri-" means three-and-a-half; and so on.
Ancient Egyptian had a special fraction for two-thirds, that wasn't just adding up unit fractions.
Guitarplayer wrote:
me wrote:You have multiplicative numbers like twice and thrice etc.; can you also have numbers that multiply by a fraction?
I don't see how the last question is practical (how often do you say "three-quarter-times"?). It would be fun to figure out, though.
In English we often multiply by one-and-a-half; "half again" is a common phrase. Maybe I'm wrong but I think we also often multiply by two-thirds. And we decimate (we learned how from the Romans' Latin); that's basically multiplying by nine-tenths.

If you have words for:
unit fraction
one minus a unit fraction
one plus a unit fraction
you may, or may not, have words related to those terms the way "once" is related to "one", "twice" is related to "two", and "thrice" is related to "three".

You might also have a series of words for
integer plus a half
and you may have some words for multiplying by those.

But in general I expect
integer plus unit fraction
integer minus unit fraction
integer times unit fraction
to be phrases rather than words, except in special cases (at the outside, when the integer and the denominator are both less than or equal to your base; though maybe in some cases you'll allow the unit fraction's denominator to be some power of your base).

_____________________________________________________________________________________

Looking forward to whatever comes next!