qiihoskeh: myo: kanji (Default)
* I discovered a couple more problems.

11. Embedded WH-Questions and Relative Clauses
An embedded WH-question (and the normal variety of non-embedded WH-questions, for that matter) begins with an interrogative pronoun; this is either koi or a word form beginning with the k- morpheme. A relative clause begins with a relative pronoun, which begins with the y- morpheme. Often, this isn't a problem, due to the fairly free order of phrases and phrase components and because required possessors/objects precede their heads. Here are some examples (using non-embedded questions for simplicity).

Koi hote tafo java? -- "How hot is your coffee?"
Kanok ta vid'he? -- "What (thing) have you seen?"
Kamak dok zan disek? -- "Whose mother told them that?"
Kilko hauso cabi korek zo hundak? -- "What kind of house did the dog run to?"

The problem occurs when the interrogative or relative pronoun would be contained within a subordinate clause. English pulls the pronoun (or the phrase containing the pronoun) out of the subordinate clause, as in "What do you want us to give you?" where "what" is really an object of "give", not "want" (Note: the English infinitive construction corresponds to an MNCL5 complement clause). I don't see this working for MNCL5. Possibly some kind of "such that" construction could be used, with perhaps a 3rd person pronoun remaining in place; I haven't figured out the other details yet. I wonder what other strategies might be used?

12. Order of Medials
I've tried to arrange things so that all inflectional medials follow all derivational morphemes. However, I've come across a possible exception. It seems to make more sense for -apt-, which means "likely to" and is clearly derivational (it changes the argument structure), to be applied to a stems ending in -es-, which patterns with the aspect medials, as in:

Marxok disesapte. -- "Marsha is talkative." (literally, "likely to continue talking")

Another possible exception is mentioned below.

* Regarding the earlier problems, I've made a little progress.

7. Tetravalent Verbs
I think the morpheme used for the stake or price (which may be uxc-) could be translated as "offer" or "tender", in which case it relates to the seller, bettor, etc. This means that the -i final should be used, as in:

Jonak Toma vendek horsa pento gat'luxci.
"John sold Tom a horse for five cats."

3. Trivalent Imperatives
5. Conatives

I haven't made a decision on these yet, but I think they're related: both involve making a distinction between two arguments. So, either the -m- or the -s- could be used with the conative marker, whichever is used with the imperative. This brings up another issue, however: must the conative medial precede the grammatical voice medial or can it follow?

Purpose in MNCL5

2007-Dec-27, Thursday 02:31
qiihoskeh: myo: kanji (Default)
I think I found some more problems with MNCL5 and solved one of them before I even realized it was a problem. Here it is, with some background first.

Originally, MNCL5 had a simple rule stating that omitted arguments (i.e. those which are allowed grammatically but don't appear) are, if anything, indefinite (Note that this is distinct from the implied subjects of non-verb forms). Next, when I was figuring out how to express purpose in MNCL5, I came up with the idea of using an adverbial complement clause containing a subjunctive verb, as in:

Mas kor'yek gil Jonita zo elefantok vidus di.
"We were running so that Johnny could see the elephant."

For some of the cases where the complement clause shares a referent with the matrix clause, I decided that a prospective aspect participle component with a secondary predicate final would be used, as in:

Jonitak korek zo elefanto vid'vi.
John-DIM-AGT.SG run-PST 3A-ADJ elephant-GEN.SG see-PRO-ADV
"Johnny ran to see the elephant."

(In this example, the secondary predicate final -i is identical to the adverbial final, and has been glossed the same, even though the semantics are different)

However, this leaves some cases where a referent is shared, but the clause can't be replaced by a component (such as when it has too many arguments or an adverbial pharse). I originally used a 3rd person pronoun, or maybe a reflexive pronoun, in the complement clause for the shared referent, but later decided it could be omitted in many cases, even though this would require the omitted argument rule to be modified. Example:

Jona cuzek Tomok gil (zak) Marxa gebus livrok din.
John-PAT.SG choose.PST Tom-THM.SG CCI (3A-AGT.SG) Marsha-PAT.SG give-SUB book-THM.SG SCT-SEC
"John chose Tom to give Marsha a book."

Note that -in (SEC) is used instead of -i (ADV), since the coreferenced argument is Tomok, not Jona.

Recently, I realized that ambiguity could occur if an additional argument were omitted [Assume that there's an adverbial phrase preventing the clause from being replaced by a component].

Jona cuzek Tomok gil gebus livrok [tomorrow] din.
John-PAT.SG choose.PST Tom-THM.SG CCI give-SUB book-THM.SG [tomorrow] SCT-SEC
"John chose Tom to give/be given a book tomorrow."

A couple things could be done: (a) restore the 3rd person pronoun or (b) use an indefinite pronoun for the other omitted argument. But then I came up with something I like better: use a relative clause instead of an adverbial clause. This uses one word fewer than (a) or (b) and is actually more consistent with the grammar. Example:

Jona cuzek Tomok yak gebus livrok [tomorrow] din.
John-PAT.SG choose.PST Tom-THM.SG REL-AGT.SG give-SUB book-THM.SG [tomorrow] SCT-SEC
"John chose Tom to give a book tomorrow."

This also eliminates the need to modify the omitted argument rule!

MNCL5 Problems

2007-Dec-25, Tuesday 01:41
qiihoskeh: myo: kanji (Default)
MNCL5 has accumulated a large number of unsolved problems. I'm going to try to list all the current ones here, so they'll all be in one place.

Note: "medial" is my term for "non-final suffix".
Read more... )
That's all I've found in my notes, except for a neverending list of lexical decisions.

(no subject)

2007-Dec-22, Saturday 04:24
qiihoskeh: myo: kanji (Default)
I've been thinking a bit how I might make a conscenario for MNCL5, to "justify" its peculiarities.

To review, MNCL5 is not very naturalistic; it's very regular, there's little or no lexical compounding, all non-initial morphemes begin with vowels, the syntax uses embedding, most roots are borrowed (it's easier to invent and remember words this way), and maybe a couple other things that escape me at the moment. At the other end, it's not a loglang, definitely not a taxonomic language, and not like Ithkuil either; also, it's not designed to be an IAL. So, all this places limits on the scenario.

What I'm considering is this: at some time in the past, a group of aliens came to Earth and transformed themselves into humans, albeit with some subtle differences, mainly in the brain I guess, so that their fundamental character was preserved. Somehow (I need to figure this out), their alien origins were forgotten, so that the present-day descendents think they're purely human. By chance, a group of them come together. Dissatisfied with human languages, because their brains aren't perfectly adapted to them, they create a private auxlang (which is of course MNCL5). I think this explains the peculiarities of the language, although there a lot of details to be worked out. For instance, how long ago the aliens arrived and whether they can breed with original humans or constitute a parallel species.

For fictional purposes, I may add the following. The aliens were followed by aliens of a second kind who were persecuting the first kind. The second group, who have also transformed themselves into humans, also retain their fundamental character. They basically have no redeeming qualities, but are clever in disguising this. The first group tend to be gentle and pacifistic except that they can become instinctively violent when encountering the second group.

(no subject)

2007-Nov-29, Thursday 06:28
qiihoskeh: myo: kanji (Default)
* I tried writing up MNCL5 Adverbs and Secondary Predicates and Usage of Tenses and Aspects, but got bogged down pretty quickly. For the former, I need a good explanation of the difference between adverbs and secondary predicates. For the latter, I have some handwritten notes, but these aren't sufficient for a coherent explanation.

* I've read the third chapter of the dissertation I mentioned in the previous entry. There's a bit of theoretical stuff in the beginning that I didn't quite get involving semantic features, but the rest is easy. The author examines variations on a theory correlating Slavic aspect with Germanic definiteness and effectively demolishes it. As a native speaker of English, it seemed obvious to me that such a theory wouldn't work.

(no subject)

2007-Nov-26, Monday 03:19
qiihoskeh: myo: kanji (Default)
* I've had a cold for the past couple weeks and have gotten very little done except for work on MNCL5. For that, I've worked mainly on syntax, but also on morphology, and have written down (but not typed up) a list of possible vocabulary items. There are still some details of phonology that need to be decided, such as whether p, t, k, and c are aspirated enough to mention it and how some clusters are pronounced, including tr, dr, ts, dz, cr, and jr. I probably also should list which two-consonant sequences don't occur. These include all double consonants plus tx, tc, dj, and dy.

* The last couple of days, I've been reading a dissertation on Polish verbs. So far (the first two chapters) it's been very readable (except for the German quotes) and non-theoretical. Chapter 1 gives a clear overview of Polish verb pairing and an outline of the historical developments, starting in Proto-Indoeuropean times. Chapter 2 discusses the history of the terms "aspect" and "Aktionsart".

* The only thing special I did for Thanksgiving was call my sister.

* I still need to log my MRI adventure. I hope I still remember everything by the time I get around to it.

MNCL5 "Participles"

2007-Nov-14, Wednesday 15:27
qiihoskeh: myo: kanji (Default)
MNCL5 non-verb forms are sort of like participles in some other languages, only the same general system is used regardless of whether the stem is a verb-, adjective-, or noun-type stem. Of course, the noun- and adjective-stem forms aren't normally considered participles! They're mentioned below for completeness.

There are four types of participles in MNCL5. The 1st is constructed without adding any special suffix. The others are constructed with the following suffixes:
2nd: -m- suffix
3rd: -t- suffix
4th: -g- suffix

Which of these can occur depends on the word stem's argument structure. Whether or not the participle must be preceded by an "object" (in the genitive case) also depends on the argument structure.
Read more... )

(no subject)

2007-Nov-12, Monday 13:29
qiihoskeh: myo: kanji (Default)
* I haven't gotten much done the last couple days; I've felt mostly tired and my concentration hasn't been good. It's annoying to have plenty of time and not be able to use it.

* I did work some more on the MNCL5 morphology (Saturday ?) and started working on the syntax by copying the MNCL4 file and changing a few things. I also found the diskette with the Vallés files and loaded them. I've forgotten so much.

(no subject)

2007-Nov-05, Monday 00:09
qiihoskeh: myo: kanji (Default)
* I've still been sleepy a lot. I was going to drag myself out to the grocery store tonight but fell asleep while sitting up on the couch. I'm supposed to go through my clothes and find the ones I can't wear anymore, bu I haven't been able to get started.

* I took a break from classifying compound sub-types to work on a couple of other things. I copied the file containing the list of MNCL4 medials, deleted the obsolete ones and added the new ones from my MNCL5 notes. I have to somehow distinguish the true medials from the vowel-initial initials (to complicate things, some morphemes are both). These are all long medials; the short ones will probably be covered elsewhere, since they're mostly voice and aspect markers. I also started a TOC file. I need to write up the general morphology, but can't decide if I should describe the morpheme types orthographically or phonologically. I also need to write up the phonology and orthography, which aren't so simple anymore.

(no subject)

2007-Nov-03, Saturday 07:22
qiihoskeh: myo: kanji (Default)
* Thursday, I got the referral to the neurologist; the appointment is on Tuesday and in the same building. Afterward, I went to the MBHC drop-in center (my "club") just in time for lunch. The authorization finally came in, so I have an appointment with my therapist next Thursday. When I got home, I slept for several hours, since I hadn't slept much at night.

* Since then, I've actually been conlanging for a change: specifically, working on MNCL5 syntactical compounds (there's no morphological compounding, only derivation). First, I downloaded sections 14 and 15 of chapter 5 of the Lojban Reference Grammar and did some editing to make them more readable. They contain detailed descriptions of each of the semantic sub-classes of lojban "tanru". Then I copied them into the file which will eventually be used for documenting MNCL5 compounds. Now I've started adding notes about how each sub-class will be implemented in MNCL5. Once that's done, I'll rearrange the sections according to the kind of implementation (I'll have to explain the different kinds elsewhere). Finally, once I've added enough of my own examples, I'll eliminate the lojban documentation from the file. I expect this process will take a while.

* Yesterday, I slept most of the day, and have taken three naps since, so I haven't gotten anything done except work on MNCL5. I still need to take a shower and write some emails.
qiihoskeh: myo: kanji (Default)
Argument Structure

Verb Forms

1. typical transitive verb stems: the agentive is used for the agent and the patientive is used for the patient.

Jonak Toma poncek. - "John hit Tom."

2. ditransitive/trivalent verb stems: the agentive is used for the donor, the patientive is used for the recipient, and the thematic is used for the 3rd argument.

Jonak zo cika mexek librok. - "John gave the boy a book."

3. activities: the agentive is used for the actor.

Jonak kore. - "John is running."

4. other divalent verb stems; this includes spatial relations: the patientive is used for the perceiver etc. and the thematic is used for what's perceived etc. For some of these, the patientive can be replaced by an agentive with change of meaning.

Jona videk zo katok. - "John saw the cat."
Jonak videk zo katok. - "John looked at the cat."
Zo kata no husok ce. - "The cat is at the house."

5. relational noun stems: the patientive is used for the possessum and the thematic is used for the possessor; the verb af- "belong to" also follows this pattern.

Jona Tomok atune. - "John isn't Tom's father."
Zo hunda Jonok afe. - "The dog belongs to John."

6. other monovalent verb stems: the patientive is used for the subject. For some of these, the patientive can be replaced by the agentive.

No husa ruye. - "The house is red."
Jona dorme. - "John is sleeping."
Jonak falek. - "John fell (intentionally)."

7. monovalent noun stems: the patientive is used for the subject.

Jona katune. - "John is not a cat."

8. impersonal verb forms: no arguments are used.

To complicate things, some verbs can take additional arguments, but this isn't well-defined yet.

Non-verb Forms

Non-verb forms have a "subject", which is not expressed, and, if not monovalent, an "object" (or possessor), which takes the genitive case.

1. typical transitive verb stems: forms not marked for voice are "passive" (i.e. the subject is the patient and the object the agent). The -m- medial is used for "active" forms (i.e. the subject is the agent and the object the patient).

2. ditransitive/trivalent verb stems: if the form isn't marked for voice, the subject is the recipient; the object can be either the donor or the 3rd argument. If the -m- medial is used, the subject is the donor; the object can be either the recipient or the 3rd argument. If the -t- medial is used, the subject is the 3rd argument and the object is the recipient. If the -g- medial is used, the subject is the 3rd argument and the object is the donor.

3. activities: the subject is the actor.

4. other divalent verb stems: if the form isn't marked for voice, the subject is the perceiver etc. and the object is what's perceived etc. If the -t- medial is used, the subject is what's perceived etc. and the object is the perceiver etc. The -m- medial changes the meaning, as the agentive case does for verb forms.

5. relational noun stems: if the form isn't marked for voice, the subject is the possessum and the object is the possessor. The -t- medial inverts the roles.

6. other monovalent verb stems: the subject is the subject. The -m- medial changes the meaning, as the agentive case does for verb forms.

7. monovalent noun stems: the subject is the subject.

8. impersonal verb forms: there are no non-verb forms, unless an argument is added (in which case one of the voice medials is used).
qiihoskeh: myo: kanji (Default)
Note: the orthography used is the short version, which omits predictable vowels, gemination, etc.

Aspect and grammatical voice are marked by medials (non-final suffixes). While voice marking occurs only on non-verb forms, aspect marking can occur with almost any final.

The aspect medials for dynamic verb stems are:
-y- - imperfective
-h- - retrospective
-v- - prospective
with perfective being unmarked. An exception to this is that the imperfective marker is usually omitted with present tense.

Static verb stems don't take the imperfective medial, being naturally imperfective. Another thing to note is that aspect marking doesn't occur with the tenseless final, and usually not on noun stems.

The voice medials are:
-m- - agentive subject (patientive or thematic or no object)
-t- - thematic subject (patientive object)
-g- - thematic subject (agentive object)
with patientive subject (agentive or thematic or no object) being unmarked. The word's argument structure determines which of these can occur.

When both occur, the aspect medials appear first.

There are two other medials that can occur before the voice medials and that derive noun stems; -z- for animate and -n- for inanimate.

Finally, there are other aspect-related medials such as the habitual marker -oft-.

Ta hoskvuke zo sailoftmok? "Have you heard the singer?"
2-Pat.S hear-Pro-Int-Prs Def-Adj sing-Habt-AgV-Thm.S
[ta: hOs.ku"vuk:e: zo: sai"lOf.tO.mOk]

MNCL5 Finals

2007-Sep-16, Sunday 19:19
qiihoskeh: myo: kanji (Default)
The MNCL5 final suffixes are divided into verb-form finals and non-verb-form finals. The verb-form finals are:
-e - present tense (indicative mood)
-ek - past tense (indicative mood)
-es - future tense (indicative mood)
-en - tenseless (indicative mood)
-u - imperative mood
-us - subjunctive mood
-uk - contrafactual mood

The non-verb-form finals are divided into noun finals (case-and-number markers) and other suffixes. The noun finals are:
-a - absolutive singular
-an - absolutive plural
-ak - ergative singular
-as - ergative plural
-ok - thematic singular
-os - thematic plural
-o or 0 - genitive singular
-on or -l - genitive plural
-ai - vocative singular
-au - vocative plural
"Absolutive" and "ergative" are carried over from previous versions of MNCL and aren't strictly correct. I need to change them, and I'm also not sure about "thematic". The first kind of genitive endings are used before consonants and the second before vowels (in which case the gentitive word combines with the following word).

The other finals are:
-o - adjective (actually any non-final word of a phrase)
-is - infinitive
-i - adverbial
-i - secondary predicate, type 1
-in - secondary predicate, type 2

There are also particles ending in -l.

(no subject)

2007-Sep-14, Friday 22:07
qiihoskeh: myo: kanji (Default)
* Since last Friday, I've been working on MNCL5, the newest version of an old project. It's complicated and not stable yet, so I'll just give a bit of the history.

MNCL wasn't so complicated. It had strictly CV syllables and three types of morphemes:
roots or "initial" morphemes of the form C- or CVC-,
non-final suffixes or "medial" morphemes of the form -VC-, and
final suffixes or "final" morphemes of the form -V.

The finals determined how the word functioned syntactically:
-e made the word a verb,
-i made the word adverbial,
-a made the word a noun with absolutive case, and
-u made the word a noun with ergative case.
-o was used if the word wasn't the last word of a phrase; this could mean that the word was an adjective or a noun with the 3rd case.

Actually, "noun" and "adjective" here are probably misleading, since the noun wasn't necessarily a lexical noun, nor the adjective a lexical adjective (i.e. a static verb to which comparison could be applied.)

The basic system has been kept except that diphthongs and clusters were introduced at some point. MNCL5 allows some final consonants, but has dropped onset clusters. An addition which impacts the orthography is the introduction of short medial syllables. These are written as single consonants, with the preceding short vowel determined by its context.

The final consonants and diphthongs allow tense and mood to be marked by the finals and for number to be fused with case. The aspect and voice suffixes have become short medials instead of long ones. The syntax hasn't changed much; infinitives have been added, and the normal word order, which wasn't really fixed, will probably change.


qiihoskeh: myo: kanji (Default)

August 2017

2122 2324252627


RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated 2017-Sep-22, Friday 02:47
Powered by Dreamwidth Studios