Alál: Oh Look, More Verb Things
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Alál: Oh Look, More Verb Things
Introduction
Figured it was time I returned to Alál.
This is an agglutinative a priori language that makes great use of infixation (and inininfixfixfixation), reduplication, and admittedly overly-complicated verbs. It has an attached conworld and conspeakers, who have human-like but nonhuman cognition; I have a tendency to go off on tangents and ramble on about unrelated or tangentially related topics regarding this conworld and its inhabitants' cultural practices, and will put these rambles under cuts wherever possible. My apologies for that in advance. There are likely to be many things changing as I go, particularly as some pages of my entirely unsorted scribbles may contradict other pages. I greatly welcome comments, criticism, critique, questions, and all of those good things. This mess isn't intended to be particularly naturalistic, but it is supposed to have a decent amount of internal consistency.
Figured it was time I returned to Alál.
This is an agglutinative a priori language that makes great use of infixation (and inininfixfixfixation), reduplication, and admittedly overly-complicated verbs. It has an attached conworld and conspeakers, who have human-like but nonhuman cognition; I have a tendency to go off on tangents and ramble on about unrelated or tangentially related topics regarding this conworld and its inhabitants' cultural practices, and will put these rambles under cuts wherever possible. My apologies for that in advance. There are likely to be many things changing as I go, particularly as some pages of my entirely unsorted scribbles may contradict other pages. I greatly welcome comments, criticism, critique, questions, and all of those good things. This mess isn't intended to be particularly naturalistic, but it is supposed to have a decent amount of internal consistency.
Last edited by kiwikami on 21 May 2019 18:13, edited 16 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
Re: Lılalkìtakálxıtlál (Alál)
Condangdernit, what a lay-out! КϒΔΟΣ!
The minuciosidad painstaking attention to detail in your sections... the humor... awe-inspiring!
A review of some of my faves:
Table of Discontent -
the rebel-yell of all langers and glossopoets
Diphthongs Probably Don't Work Like This But That's Fine - a.k.a. Imok UR a diphthong
http://i2.wp.com/www.drthomasharris.com ... =462%2C614
Noun Cases, Augmentatives, and Diminuitives: The Musclefluffball Children -
؟ Musclefluffball ‽
Aspect and Permanence: Telicity Gets Mingled With Everything And I Don't Know Anymore -
a.k.a. Funkentelicity vs. the Placebo Syndrome
Relationship with Zta - KNEEL BEFORE ZTA!
http://www.fansshare.com/media/content1 ... p-8713.jpg
Frequentative: Reduplicative Infixation Strikes Again - I'm looking forward to this because (1) I reduplication and (B) I'm looking forward to this sequal to Inceptive and Inchoative - Every New Hope Awakens...
Arguments of Participles: Kxa 2 - The Return of Kxa - Big Fan of PTCPs, but especially...
WH-Questions with tt : DRWH-Questions with ttaarrddiiss
Color Terms pfffft who am I kidding, it's been a year and I still have no idea how these'll work
The humility. The humor. The Thoughfulness. The Planning. [impressive. most impressive.]
Diggity Danguage
Cannot wait to see more.
The minuciosidad painstaking attention to detail in your sections... the humor... awe-inspiring!
A review of some of my faves:
Table of Discontent -
the rebel-yell of all langers and glossopoets
Diphthongs Probably Don't Work Like This But That's Fine - a.k.a. Imok UR a diphthong
http://i2.wp.com/www.drthomasharris.com ... =462%2C614
Noun Cases, Augmentatives, and Diminuitives: The Musclefluffball Children -
؟ Musclefluffball ‽
Aspect and Permanence: Telicity Gets Mingled With Everything And I Don't Know Anymore -
a.k.a. Funkentelicity vs. the Placebo Syndrome
Relationship with Zta - KNEEL BEFORE ZTA!
http://www.fansshare.com/media/content1 ... p-8713.jpg
Frequentative: Reduplicative Infixation Strikes Again - I'm looking forward to this because (1) I reduplication and (B) I'm looking forward to this sequal to Inceptive and Inchoative - Every New Hope Awakens...
Arguments of Participles: Kxa 2 - The Return of Kxa - Big Fan of PTCPs, but especially...
Spoiler:
Color Terms pfffft who am I kidding, it's been a year and I still have no idea how these'll work
The humility. The humor. The Thoughfulness. The Planning. [impressive. most impressive.]
Diggity Danguage
Cannot wait to see more.
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Lılalkìtakálxıtlál (Alál)
Thank you.Lambuzhao wrote:[many things!]
A brief explanation:Lambuzhao wrote:Noun Cases, Augmentatives, and Diminuitives: The Musclefluffball Children -
؟ Musclefluffball ‽
Spoiler:
Consonant Phonology, Orthography, and Whatnot
"S/ᵑkʼ/ooby-Dooby-Doo, What Are You?"
Behold. Sounds.
/t̪ ʈ k/ ⟨t d k⟩
/t̪ʼ ʈʼ ᵑkʼ/ ⟨ṭ ḍ ḳ⟩
/s̪ ʂ ɬ ç x~χ/ ⟨s z l h x⟩
/t̪͡s̪ ʈ͡ʂ t͡ɬ/ ⟨ṣ ẓ ḷ⟩
/ɾ~ɺ/ ⟨r⟩
/i~e~ɛ~j u~o~ɔ~w æ~ɑ~ɐ~ʕ̞/ ⟨ı a u⟩
I'll discuss only the consonants here. The vowels are complicated and get their own post.
A few liberties were taken in the creation of this phonology (which is, in-universe, partially artificial), and there are a couple of things described below regarding allophony in Lılalkìtakálxıtlál (the proto-lang, spoken by nonhumans) that become rather opaque after the modality switch (to a human articulatory system). So when ejectives are prenasalized intervocalically and some root-final /r/s after /ɑ/ trigger vowel height changes, know that while there is some semblance of internal consistency, one should keep in mind that the language was not originally spoken, and that the original auditory-modality variant was essentially a cipher, created by a small handful of human and nonhuman speakers. It has since evolved, but the "Pre-Terran" variety described here is that of the sixth-ish generation of native (or human-L2) bilingual Lılalkìtakálxıtlál-[some human language] speakers.
Just... keep in mind this isn't exactly supposed to be naturalistic.
/t̪ ʈ k/ ⟨t d k⟩
Note that while every oral stop has a corresponding fricative in its place of articulation, there are both palatal and velar~uvular fricatives (and a lateral, which has no oral stop at the same POA). This is a case of Lılalkìtakálxıtlál-to-Alál not matching up perfectly; in Lılalkìtakálxıtlál, the lateral fricative equivalent is actually the one that corresponds most closely to the phoneme that here becomes /k/ (thus the presence of phonemic /t͡ɬ/ but not /k͡x/). The "velar" is actually underlyingly the uvular /χ/, as evidenced by its effect on adjacent velar sounds, but it typically surfaces as velar elsewhere.
(t̪,ʈ,k)ç > t̪ʲʰ,cʰ,kʲʰ
C[velar] > C[uvular] / _χ
/t̪ʼ ʈʼ ᵑkʼ/ ⟨ṭ ḍ ḳ⟩
These were clicks once, /ǀ~ᵑǀˀ~ǀʰ/ /ǃ˞~ᵑǃ˞ˀ~ǃ˞ʰ/ /ᵑǃ͡q~ɴǃ͡ɢ~ǃ͡qʰ/, with intervocalic "glottalization" and "nasalization" due to modality quirks of Lılalkìtakálxıtlál translating less-than-perfectly. These clicks remained only for the nonhuman Lılalkìtakálxıtlál speakers who developed Alál (and a subset of linguistically able human speakers); they were reanalyzed as ejectives by early generations of L2 human speakers, most of whom learned from the human speakers who were not in that subset and did not natively speak a language with clicks. The intervocalic nasalization remained, with the already-nasalized alveolar-uvular contour click (recall from the fricatives how the uvular and velar POAs are somewhat fluid) eventually becoming a pure velar nasal.
t̪ʼ ʈʼ ᵑkʼ > ⁿt̪ʼ ᶯʈʼ ŋ / V_V
/s ʂ ɬ ç x/ ⟨s z l h x⟩
ɬ > ɭ \ʈ_
(s̪s̪,ʂʂ,ɬɬ,çç,xx) > (s:,ʂ:,ɬ:,ç:,x:)
Retroflex consonants are subapical palatal; these are pronounced with a very concave tongue, and produce r-coloring on preceding vowels (think Dravidian languages). Dental+retroflex/retroflex+dental clusters other than /ʂs̪/ and /s̪ʂ/ are disallowed. If the retroflex component is a fricative and the dental one a stop, the phonemes meet in the middle; the dental one moves back and the retroflex one forward, though maintaining its palatal component. If the retroflex is a stop, regardless of the dental, it becomes an approximant while the dental sound moves back:
t̪ʂ > tʃ
s̪t̪ > ʃt
s̪ʈ > sɻ
ʈs̪ > ɻs
ʈt̪ > tɻ
t̪ʈ > ɻt
/t̪͡s̪ ʈ͡ʂ t͡ɬ/ ⟨ṣ ẓ ḷ⟩
There's not much to say about these as of yet, save that affricates contrast with stop+fricative sequences at the same POA.
(t̪͡s̪t̪͡s̪,ʈ͡ʂʈ͡ʂ,t͡ɬt͡ɬ) > (t̪t̪͡s̪,ʈʈ͡ʂ,kt͡ɬ)
/ɾ~ɺ/ ⟨r⟩
This is a tap (more of a trill in Lılalkìtakálxıtlál) that for the most part assimilates to the POA of the preceding consonant, though it is alveolar word-initially and intervocalically and thus could be said to be underlyingly /ɾ/.
ɾɾ > r
There are also a pair of phonemes consisting of /ɾ/ plus a high or low floating... something (there used to be tones - there aren't anymore) that attaches to a preceding mid-tier vowel, written as ⟨ŕ⟩ and ⟨r̀⟩ in glosses but not distinguished in the surface orthography. Lılalkìtakálxıtlál used to distinguish between three trill-equivalents at high, middle, and low frequencies just as with all other phonemes, but the distinction was lost in most roots, collapsing into /ɾ/ except after the middle-frequency phoneme that is realized in Pre-Terran as /æ~ɑ~ɐ~ʕ̞/. This was to avoid the two similar sounds becoming conflated, but the reasoning for this is rather opaque in Pre-Terran, where there's not much of a connection between /æ~ɑ~ɐ~ʕ̞/ and /ɾ/. The result, anyhow, is that while these /ɾ/s are still found in plenty of affixes and non-root morphemes (such as the general plural subject marker -ŕ-), you only see them in a handful of roots, always as the second consonant, and always after the root vowel /ɐ/.
V(r̀,ŕ) > (V̀,V́) *
(r̀,ŕ) > ɾ **
The rule marked * occurs before initial vowel epenthesis (described below), while the rule marked ** occurs afterwards. Thus, kıŕsı > kísı > ıkísı /ɛ'ki.z̪ɛ/ “they are being born”.
Regarding Clusters
All two-phoneme clusters are permitted in any position, with the exception that stop+stop clusters of the same POA may not appear initially. Clusters of three consonants are not permitted word-initially or finally, and are only permitted medially if all three are not identical. Should identical CCC clusters occur through affixing, the vowel of the most salient root (see mention of initial vowel epenthesis above) is epenthesized between the first and second elements. Similarly, if four-consonant or higher clusters should occur, that vowel is epenthesized before the second-to-last element: thus, CCCC > CCVCC and CCCCC > CCCVCC.
Alál is built mostly on C(C)VC roots. These roots will (
Some Notes on Orthography
In the Latin-alphabet orthography, verbs (not participles or nominalized verbs) are capitalized, as are the first words of a sentence. Foreign words, as well as words being mentioned rather than used – essentially any nouns that cannot be declined for case or don't have a clear 'grade' (I'll discuss noun class/grade later) – are placed within interpuncts. Direct quotations, as well as longer phrases in a foreign language, are placed between colons. When to use interpuncts vs. colons is related to when to use different grades of the complementizers/discourse markers/case-taking proforms raŕ and das, which I'll get to later. There's a bit of a dot-heavy aesthetic going here, in case you couldn't tell. I like dots.
A Final Thing Regarding Voicing
Intervocalic voicing occurs before unstressed syllables:
Hısıl ['çe.z̪ɛɬ] 'he flinches'
Ktaḷı ['ktɑ.d͡ɮɛ] 'I remember you'
More to come as I remember everything I'm certain I've forgotten...
Last edited by kiwikami on 21 May 2019 18:28, edited 10 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Lılalkìtakálxıtlál (Alál)
Edit: When more than one example is given, it is now put under spoiler text for de-cluttering purposes.
Roots, Word Order, and Basic Clause Structure, and ThingsHere, I'll give a relatively concise (lies. all lies.) summary of the structure and functions of Alál's roots, its basic word and constituency order, basic clause structure (not going to deal with relative clauses or how participle constructions work just yet) and examples of exceptions and small oddities, namely, the few instances where the verb is not the first element of a clause and what they entail (sentence-initial agentive nouns as vocatives, sentence-initial oblique nouns as a variety of constructions related to focus, possession, and membership in a set, and sentence-initial patientive nouns in WH-questions), along with the rare instances of VSO and some stuff about volition. I'll also mention some things about serial verbs (and nouns), and briefly talk about how prepositional phrases work, though I'll get into that in far more detail later when I actually list the prepositions. This'll be spread through a couple of posts so I can link to the topics separately in the Table of Discontent.
Again, at any given point in time it is safe to assume that I have no idea what I'm doing, and so comments/critique and extremely welcome.
Roots and Things
Alál is built primarily on C(F)VC roots, each of which indicates some usually small set of semantic meanings; each root may serve as a noun in one of three magnitudes, a transitive or intransitive verb in one of ten (or thereabouts?) aspects, or as part of several different types of compound. The relationships between various uses of a root can rarely be assumed; though the meanings are usually related, they cannot always be derived from each other in a logical pattern, and the meanings of each root in various positions and as various parts of speech typically must be learned lexically. Given transitive vs. intransitive verb forms, for example, one cannot assume that the former may be a causative form of the other (this is only very rarely the case).
Examples of this disparity in meaning include the root hus, which as a noun refers to a flame and as a transitive, durative verb may be translated as "to burn (OBJ - a fuel)", but which as an intransitive, momentane verb means "to jerk suddenly" or "to flinch". Another root, kıs, is used as a noun to refer to Wayfarers in one of the three juvenile life-stages (pre-sexual maturity), or by humans to refer to infants, toddlers/small children, and preteens, depending on magnitude. As an intransitive verb, it generally means "to be born" or "to be young" depending on aspect; as a transitive verb, it can depict a wide variety of familial relations. When used as the secondary root in a noun compound, however, it invariably refers to unformed or incomplete versions of the primary root, where magnitude can be used to indicate extent of completion (examples given in the oblique case):
ıkásuhas kıs (incomplete) (AUG) + hus (fire) = unlit kindling
ıkàısuhas kıs (incomplete) (DIM) + hus (fire) = smoking/smouldering material with no visible flame
kasıtxáır kıs (incomplete) + txır (sphere) (AUG) = dome (as used in architecture)
ıkáısıtxàır kıs (incomplete) (AUG) + txır (sphere) (DIM) = a small, curved solid object, such as a piece of a broken lightbulb (Essentially anything resembling a piece of an eggshell. Except an eggshell. Eggs are special. They have their own root, ḍax.)
kasıtxàır kıs (incomplete) + txır (sphere) (DIM) = bowl or bowl-shaped small object
ıkàısıtxar kıs (incomplete) (DIM) + txır (sphere) = medium-sized roundish thing with one or a few small openings (Likely either "bowling ball" or "human skull" depending on who you ask; a third root can be added if you need to be specific about it, but the situations in which those two things could be confused are few in number. If you made txır diminutive too, you could potentially have "olive". Maybe also "pincushion" and "tennis ball stuck on the end of a pole to make the pointy end less pointy, or on the leg of a chair to make it not scrape the floor" out of it.)
But I'll ramble on about that more when I actually get to a section devoted to compounding. There's a lot of compounding.
Word Order and Things
For the most part, word order is VOS, though VSO is also common, depending on volition. Alál uses prepositions, and places participle phrases (the closest thing it has to adjectives) before the words that they modify. Each sentence typically consists of at least one independent clause, the only mandatory component of which is a verb; many clauses consist only of a verb, as the language exhibits polypersonal agreement (for subjects and objects, not counting incorporated locatives) and is pro-drop:
Lıílaı.
speak<3.3>-DUR-VOL.ACT
He is speaking with him.
Subjects of transitive verbs are in the agentive case and objects of transitive verbs in the patientive (fluid-S alignment); the marking of the subject of an intransitive verb is determined by the volition of that subject (which is indicated on the verb). Non-volitive subjects are marked as patients, where volitive subjects are marked as agents. I'll get to the specifics of this later when I talk about volition as a whole, but simply to illustrate the three behaviors of nouns in transitive vs. intransitive sentences, see below:
Spoiler:
Uksásaı kus kıh.
we_are_in_a_line-VOL.ACT child<PAT.DIM> person<AGT>
We people are lining up for the teenager.
Similarly, verbs marked as non-volitive, with patientive-case subjects, may take agentive-case nouns as well; these typically indicate some indirect causation, the means by which the action occurred, though there are also instances where they are closer to instrumental. This, and the benefactive patientive-case nouns, is all tied up with a couple of prepositions, zta and kxa, which are closely related to the case-marked nouns and which serve some important purposes in voices other than the active, but I won't talk about those yet. Below is an example of a patientive-case subject of a non-volitive verb, with an agentive-case noun along for the ride:
Hısıl ıkùs hıs.
3SG_jerked_suddenly child<PAT> fire<AGT>
The child flinched back from the fire.
All of these arguments must be either nouns declined for case, or one of raŕ or das, which carry case for arguments that cannot themselves decline for it, including entire phrases and additional clauses (the two differ in where they are located relative to the argument in question, and tend to be used in particular situations; I'll talk about them later... probably).
Serial Things
Both multiple verbs and multiple arguments marked for the same case may appear consecutively to indicate related or simultaneous actions or the presence of more than one entity acting as subject or object. Serial verbs describing related actions must share a tense conjugation and person agreement, and any following direct object and subject are considered to apply to both events. Consecutive verbs that describe events intended to be understood as separate, but which share tense and person markers, are typically separated by a longer pause, or may be broken up by the addition of some word between them (if they are part of a single utterance or relate an ongoing sequence of events, they may not be written as separate sentences).
Spoiler:
Examples are given below, but note that several constructions of this nature expressing belief, doubt, and the like are more commonly expressed through the declension of nouns for mood, rather than through what is essentially a modal verb construction. Sentences like those below show up when it is important to highlight the truth value of the statement (e.g. I think he is sleeping [but he's not]) or when the relationship between the verbs is more complicated than can be expressed with a single mood (e.g. to be happy because of something):
Spoiler:
Spoiler:
Prepositions and Discontinuous Things
Alál uses prepositional phrases that come before the words they modify, but the object of the preposition closest to the modified word is displaced to directly after the word, creating discontinuous prepositional phrases circling the words they modify rather than preceding them:
Spoiler:
Uúaḍ-ù has uús Rısıá zaz.
away_from-PREP fire into 3SG_moves water
He moves away from the fire and into the water.
Prepositional phrases may modify the objects of other prepositional phrases; this discontinuity still occurs and leads to a stacking effect:
Zaka Ḳılılà uaz ḍaḳ zaz.
on_top_of 3SG_died near rock water
It died on top of the rock by the water.
This can cause ambiguity when serial nouns are used:
Ḍıuaı Ḳılılà uaz ḍaḳ dlıáxas zaz.
far_from 3SG_died near rock cliff water
It died far from the rock that is near the cliff and the water.
It died far from the rock and the cliff that are near the water.
The correct interpretation of the sentence is generally left up to context and common sense, but the ambiguity may be resolved, if necessary, by instead using the same preposition multiple times:
Spoiler:
Last edited by kiwikami on 08 Jul 2017 18:17, edited 5 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Lılalkìtakálxıtlál (Alál)
Word Order and Volition
Though VOS is the norm, VSO word order may be used to indicate a particular volitive construction. Volition of the subject is marked on the verb; by default, a direct object is involitive, having no control over the occurring event. In an earlier form of Alál, word order was significantly freer, and volition was assigned to the last argument in the clause; in the language's modern form, the word order variation has been largely replaced by a circumstantial voice that allows the promotion of various arguments to subject position (which now takes the verb-marked volition). However, an exception remains in which VSO applies that volition to both the subject and the direct object (which in the old system would, with this word order, naturally take volition – being clause-final). In this case, the subject is still considered the instigator of the action, but the object has some say it its occurrence; essentially, the verb marks the volition of the subject, and if the object is anything but involitive, VSO word order is used to give it the clause-final volition that remains from the old system.
This'll be described in more detail when if I get around to talking about the volition system. Note that it can also work with intransitive verbs that have benefactive patientive-marked nouns, or causative agentive-marked nouns; I'll go into this later.
Sentence-Initial Patientive Nouns: WH-Not-Actually-Movement
WH-question structure in an earlier form of the language used a combination of the question clitic -tt and the verb Daàhìtı “I don't know (some information X)” as a request for information. These questions would take the following form:
Daàhìtı akùhatt Kxırsàılàùs taıt.
1SG_doesn't_know_3SG person<PAT>=Q 3SG_made_2SG_sick Q<AGT>
I don't know who got you sick.
Who got you sick?
This is still a valid construction but sounds rather archaic, though a version with a different clitic or in the pseudo-serial construction described in the post above (where the two verbs are adjacent and not separated by an intonation break, but have different subject/object agreement) works just fine as a statement of “I don't know X” without requesting that the unknown information be provided. Instead, over time, Daàhìtı was dropped, followed eventually by the -tt interrogative and its agentive counterpart. The result was a construction with an patientive-marked noun at the start acting as the subject of the following verb (which is still capable of taking a separate patientive-marked noun afterwards as a direct object).
Akùh Kxırsàılàùs ulùh.
person<PAT> 3SG_made_3SG_sick animal<AGT>
Who got the animal sick?
Though this initial noun must always be interpreted as if it were the agentive-marked argument of the following verb, changes in voice using the passive or circumstantial (really two variations of the same voice, with the former marked as "transitive" and the latter not, though it still may be transitive; valency gets a bit complicated when incorporation begins to occur) allow its interpretation as that verb's direct object or as the object of some preposition, rather than its subject.
This works just fine with serial nouns:
Uḍúk utùr Tuxtídzakakı.
rock<DIM.PAT> sand<PAT> 3PL_are_being_picked_up_by_2SG_(passive)
What pebbles and sand are you picking up?
Adding another argument as subject causes that argument to be treated as if it were adjacent to/serial with the fronted one, but without being in question:
Uḍúk Tuxtídzakakı utùr.
rock<DIM.PAT> 3PL_are_being_picked_up_by_2SG_(passive) sand<PAT>
What pebbles are you picking up with the sand?
Sentence-Initial Oblique Nouns: Novel Information, Focus, Possession, and Set Membership
Oblique nouns that are not the objects of prepositions may appear at the front of a sentence, indicating a topic of discussion, introducing new information, and emphasizing the focus on that new element. There are two main uses of this construction: firstly, an oblique noun may appear initially in a sentence with no overt agentive-marked subject. In this case, the oblique noun is considered to refer to the subject (the subject agreement on the verb will match this noun) and to introduce some new information to the discourse. The key distinction between this oblique-initial construction and the use of focus particles is the novelty of the information presented:
Oblique fronting may be used to refer to entities that are clearly already in the common ground, to humorous effect or to draw attention to the novelty of the situation as a whole:
Izár Raskàaılà.
1<OBL.AUG> 1SG_just_fell_down
This person called me just fell down.
Newsflash: I just fell down.
You'll never guess what I just did (it was falling down).
The second use of this construction is when the oblique noun is not an argument of the verb but is otherwise related to the event; most commonly, it either is the possessor (alienable only – may also imply general relation) of the verb's subject, indicates some set to which the subject belongs (used in conjunction with the modifier tıḷǐ or utìḷǐ), or is a member of some set indicated by the subject (in this case, the oblique noun also takes a focus or contrastive focus clitic); these second two in particular have specific implications about the prototypicality of the set-member as a member of the indicated set. These three meanings, as opposed to the above oblique-as-novel-subject meaning, are indicated by the presence of an overt subject; examples of all three are below.
All of these work just fine with serial nouns:
Ikàıs ızák Ẓıḷaılḷ ẓılakıh.
child<OBL.DIM> 1<OBL.AUG> 3SG_just_woke_up friend<AGT>
The toddler's and my friend just woke up.
Sentence-Initial Agentive Nouns: Vocatives and Whatnot
Agentive-marked nouns may appear sentence-initially, in which case they indicate the addressee, acting essentially as vocatives. There's not much more to be said about this, though note that this can combine with both initial patientive and initial oblique nouns(though initial obliques and patientives cannot occur together) actually they can, with some interesting results, but it's two-thirty in the morning and I'm going to sleep now; in this case, the agentive noun comes first:
Though VOS is the norm, VSO word order may be used to indicate a particular volitive construction. Volition of the subject is marked on the verb; by default, a direct object is involitive, having no control over the occurring event. In an earlier form of Alál, word order was significantly freer, and volition was assigned to the last argument in the clause; in the language's modern form, the word order variation has been largely replaced by a circumstantial voice that allows the promotion of various arguments to subject position (which now takes the verb-marked volition). However, an exception remains in which VSO applies that volition to both the subject and the direct object (which in the old system would, with this word order, naturally take volition – being clause-final). In this case, the subject is still considered the instigator of the action, but the object has some say it its occurrence; essentially, the verb marks the volition of the subject, and if the object is anything but involitive, VSO word order is used to give it the clause-final volition that remains from the old system.
Spoiler:
Sentence-Initial Patientive Nouns: WH-Not-Actually-Movement
WH-question structure in an earlier form of the language used a combination of the question clitic -tt and the verb Daàhìtı “I don't know (some information X)” as a request for information. These questions would take the following form:
Daàhìtı akùhatt Kxırsàılàùs taıt.
1SG_doesn't_know_3SG person<PAT>=Q 3SG_made_2SG_sick Q<AGT>
I don't know who got you sick.
Who got you sick?
This is still a valid construction but sounds rather archaic, though a version with a different clitic or in the pseudo-serial construction described in the post above (where the two verbs are adjacent and not separated by an intonation break, but have different subject/object agreement) works just fine as a statement of “I don't know X” without requesting that the unknown information be provided. Instead, over time, Daàhìtı was dropped, followed eventually by the -tt interrogative and its agentive counterpart. The result was a construction with an patientive-marked noun at the start acting as the subject of the following verb (which is still capable of taking a separate patientive-marked noun afterwards as a direct object).
Akùh Kxırsàılàùs ulùh.
person<PAT> 3SG_made_3SG_sick animal<AGT>
Who got the animal sick?
Though this initial noun must always be interpreted as if it were the agentive-marked argument of the following verb, changes in voice using the passive or circumstantial (really two variations of the same voice, with the former marked as "transitive" and the latter not, though it still may be transitive; valency gets a bit complicated when incorporation begins to occur) allow its interpretation as that verb's direct object or as the object of some preposition, rather than its subject.
Spoiler:
Uḍúk utùr Tuxtídzakakı.
rock<DIM.PAT> sand<PAT> 3PL_are_being_picked_up_by_2SG_(passive)
What pebbles and sand are you picking up?
Adding another argument as subject causes that argument to be treated as if it were adjacent to/serial with the fronted one, but without being in question:
Uḍúk Tuxtídzakakı utùr.
rock<DIM.PAT> 3PL_are_being_picked_up_by_2SG_(passive) sand<PAT>
What pebbles are you picking up with the sand?
Sentence-Initial Oblique Nouns: Novel Information, Focus, Possession, and Set Membership
Oblique nouns that are not the objects of prepositions may appear at the front of a sentence, indicating a topic of discussion, introducing new information, and emphasizing the focus on that new element. There are two main uses of this construction: firstly, an oblique noun may appear initially in a sentence with no overt agentive-marked subject. In this case, the oblique noun is considered to refer to the subject (the subject agreement on the verb will match this noun) and to introduce some new information to the discourse. The key distinction between this oblique-initial construction and the use of focus particles is the novelty of the information presented:
Spoiler:
Izár Raskàaılà.
1<OBL.AUG> 1SG_just_fell_down
This person called me just fell down.
Newsflash: I just fell down.
You'll never guess what I just did (it was falling down).
The second use of this construction is when the oblique noun is not an argument of the verb but is otherwise related to the event; most commonly, it either is the possessor (alienable only – may also imply general relation) of the verb's subject, indicates some set to which the subject belongs (used in conjunction with the modifier tıḷǐ or utìḷǐ), or is a member of some set indicated by the subject (in this case, the oblique noun also takes a focus or contrastive focus clitic); these second two in particular have specific implications about the prototypicality of the set-member as a member of the indicated set. These three meanings, as opposed to the above oblique-as-novel-subject meaning, are indicated by the presence of an overt subject; examples of all three are below.
Spoiler:
Ikàıs ızák Ẓıḷaılḷ ẓılakıh.
child<OBL.DIM> 1<OBL.AUG> 3SG_just_woke_up friend<AGT>
The toddler's and my friend just woke up.
Sentence-Initial Agentive Nouns: Vocatives and Whatnot
Agentive-marked nouns may appear sentence-initially, in which case they indicate the addressee, acting essentially as vocatives. There's not much more to be said about this, though note that this can combine with both initial patientive and initial oblique nouns
Spoiler:
Last edited by kiwikami on 08 Jul 2017 18:24, edited 4 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Lılalkìtakálxıtlál (Alál)
~VERBS~
Episode I: The Plural Menace
Verbs exist. They do a good number of things.
The general structure is as follows. It is a mess:
ROOT#1<PERSON-PLURALITY-(ASPECT?)-OBJ_INCORP>-TRUTH_VALUE-AXIS-(COMPOUND_TYPE)-ASPECT-(ROOT#2)<VOLITION.VOICE>
Each verb requires a single root in that first position; if it is not a compound, then that is the primary root of the verb. If it is a compound, then the second root is the primary root, while the first is the secondary root. This is mostly important for determining what vowels are epenthesized to break up consonant clusters, and what is reduplicated when a morpheme calls for partial reduplication of a root. Verb compounding is tremendously important, but it'll get its own post (or five) later. The first thing to be covered here will be conjugation for person and number.
Both subject and object are marked on the verb using a single fused (fusion is just a cheap tactic to make weak morphemes stronger) infix on the first root that in all cases (somewhat unusually for infixes in Alál) replaces the root vowel entirely (though it will often show up again to force initial mid-tone, break up a cluster, or as part of a morpheme that includes partial reduplication); object plurality is included in this affix, but subject plurality is not. The persons are first, second, third, and “fourth”, where fourth is more of an indefinite third person akin to “one”, introducing new information or indicating that the argument in question isn't already in the common ground, and also to refer to less-salient third persons when there is more than one being discussed. There are also four object-less infixes that cause verbs to be considered intransitive. (This doesn't necessarily mean they have a valency of one, particularly combined with the causative and circumstantial voices, but that's a post for another time). The thirty-six total subject-object infixes are given below:
Here are some examples of them in use:
Note that the first person plural here is exclusive. In order to describe inclusive “we” or any combination of arguments, these affixes may be stacked for a maximum of two; the order of that stacking is top-left to bottom-right, following the above table - not for any real linguistic reason so much as because it makes things easier on me (which isn't a very good explanation, so I welcome alternative suggestions). If this is done, any third person markers are considered to refer to the same person, which makes for some handy reciprocal and reflexive constructions (one of two typical ways to form these).
Multiple arguments in the same case (see above posts on serial constructions) can then be used to specify what these markers refer to; they are given in the same order as the agreement markers themselves:
Subject (and object) plurality can be further specified with additional infixes falling into two major categories: quantifiers and numerals.
Quantifiers indicate the general number of arguments, relative to all arguments of that type; that is: few vs. some vs. many vs. most vs. all, as well as a general plural that does not specify. With the exception of gen. plural and all, these affixes are derived from the CVC roots used for a variety of deictic constructions (including personal pronouns), defined as the proximal, medial, distal, and extra-distal roots respectively, which allows them to take diminutive and augmentative forms that vary slightly depending on that root's vowel.
A brief note on a glossing convention of mine: a lot of morphemes in Alál include, but do not consist entirely of, partial reduplication of some other element. Where this is the case, the reduplication portion at the morpheme-by-morpheme gloss level will be separated from the rest by a dot (.) and a tilde (~) in the direction of the duplicated element, combining the Leipzig rules for reduplication and fused morphemes. Additionally, “V” here refers to the vowel of the most salient root. Which root is most salient depends on where the morpheme in question is:
Quantifiers cannot stack with each other. The vowels for the first four quantifiers are in parentheses because upon actual infixation to the root, they do not typically appear; the only reason they are included here is (1) to show the connection to the proximal through extradistal roots zık, ẓar̀, saḷ, and hız, and (2) because the presence of /i/ in those roots for few and most leads to slightly different augmentative/diminutive forms.
The few-some-many-most markers (not gen. plural or all) can take what are essentially the aug./dim. oblique noun infixes (remnants of their origins as the deictic roots) to indicate “at most” and “at least”, respectively. These are -á- and -à- respectively for “some” and “many”, but -áı- and -àı- for “few” and “most” (because the proximal and extradistal both have /i/ as a root vowel, which doesn't disappear in the augmentative or diminutive oblique; see an as-of-yet unwritten post about noun cases for charts and things regarding this). And now we get some good old ininfixfixation:
There are a lot of things about scalar implicature and the like here that I haven't yet put into words. I welcome anyone who'd perhaps like to throw out some questions about what particular constructions would mean; it'd be good to solidify this part of the system a bit more and iron out any glaring semantic wrinkles. (Semantic laugh lines?)
The second type of plural marking, numerals, is direct indication of the exact number of arguments taking part in the action. Alál allows this with single roots up to 12; beyond that, compounding is possible to produce larger or more complex numbers. The largest number it's possible to specify directly in a root is, I believe, an integer just over 9.6 * 1012 (though that's... a bit excessive). This is because Alál's number system is based on twelve integer roots (thirteen if you count zero, which I'll talk about more in a bit since it's used in negation) which can be multiplied, added, and raised to the power of each other through nominal compounding.
This, and the use of these compound numerals in plural marking, will be described later when I talk about the number system; for now, we'll stick with the basic thirteen roots 0-12. Just as with the quantifiers, the central vowel of the root is omitted unless in the augmentative or diminutive (which convey the same at least/at most meaning here), and the integers 1, 3, 4, 7, and 12 take -áı- and -àı- in those cases rather than -á- and -à-, because the roots they originate from have /i/ as a central vowel.
The numerals from 1-12 are as follows: 0 (lat), 1 (xıd), 2 (taḳ), 3 (zıṣ), 4 (ḍıṭ), 5 (ẓaḷ), 6 (tul), 7 (ṭıl), 8 (ḷuz), 9 (ḳsud), 10 (ḳaẓ), 11 (dzul), 12 (klır). These are always followed by the vowel of the most salient root. When integer plurality is used, there must also be a quantifier if an actual plural meaning is to be conveyed; typically, this is the general plural. The quantifier is infixed directly before the numeral:
Other quantifiers may also be used, but this can get rather overly-complex and difficult to parse, and it is rare in everyday speech (though not as rare in more technical contexts); note that this creates a CCCC cluster in most cases, leading to the epenthesis of the most salient vowel, which here is that of the first root, creating the more acceptable CCVCC pattern):
Using an integer without a quantifier indicates a single argument in pieces, used to show the number of members of a group that is otherwise referred to as a single entity:
It has a wide variety of other uses, many somewhat idiomatic, such as to show that one is conflicted about a topic or action; these'll probably crop up more in later example sentences:
Laàtḳalaı.
la<aà-taḳ.~a>laı
<1.3-two>talk_to
I (in two pieces) talk to him.
I talk to him (but I only kind of want to).
And then there's zero, which I gloss as NEG since that's essentially what it does. There are two methods of marking negation on a verb; one is a specific negative affix which is buddies with a couple other “truth-value” affixes (used to affirm the certainty of an event's occurrence, indicate one isn't sure whether it occurred or not, or describe hypothetical situations) and the other is the use of zero as a numeral. These two methods differ greatly in scope and I'll describe how they work in more detail later. The latter method is exemplified below:
This can technically be used in conjunction with quantifiers, but the result is for the most part nonsensical, with the exception of all:
Aẓîkaltaḷı kaıh.
ẓa<ı-ŕk.~ar̀-lat.~a>ḷ-ı ka<ı>h
sleep<3-all-NEG>-DUR person<NOM>
None of the people are sleeping (because there are no people).
A similarly humorous and/or tremendously unhelpful effect is achieved by putting it in the "at least" diminutive:
Ẓılàtaḷı.
ẓa<ı-la<à>t.~a>ḷ-ı.
sleep<3-NEG<DIM>>-DUR
At least zero people are sleeping.
I have absolutely no idea whether anyone is sleeping or not. Why the hell did you ask me this question?
The "at most" augmentative results in particularly strong negation, which is typically only used when one is very irritated, very commanding, or a very bad liar:
The last "numeral", though it is not an integer and serves a special purpose, is "Undefined", ḳat. Undefined (UNDF) essentially indicates that either (1) the speaker doesn't know how many, if any, individuals are involved, or (2) the number is irrelevant, the action incapable of consideration, or some incorrect presupposition implied. That is, the answer to “Did X occur”, where X is the verb marked with UNDF, is neither yes nor no; the truth value of the statement is not “true” or “false”. It's a useful little morpheme when you're trying to discuss logic or quantum mechanics, but also serves a purpose as a verbal shrug and, particularly in questions, as a protest to some implied statement.
It, like zero, has another form that appears in a different place and isn't treated as an integer, which I'll ramble on about later.
In the meantime, at last, we'll move on to object plurality.
Later.
In another post.
Eventually.
Episode I: The Plural Menace
Verbs exist. They do a good number of things.
The general structure is as follows. It is a mess:
ROOT#1<PERSON-PLURALITY-(ASPECT?)-OBJ_INCORP>-TRUTH_VALUE-AXIS-(COMPOUND_TYPE)-ASPECT-(ROOT#2)<VOLITION.VOICE>
Each verb requires a single root in that first position; if it is not a compound, then that is the primary root of the verb. If it is a compound, then the second root is the primary root, while the first is the secondary root. This is mostly important for determining what vowels are epenthesized to break up consonant clusters, and what is reduplicated when a morpheme calls for partial reduplication of a root. Verb compounding is tremendously important, but it'll get its own post (or five) later. The first thing to be covered here will be conjugation for person and number.
Both subject and object are marked on the verb using a single fused (fusion is just a cheap tactic to make weak morphemes stronger) infix on the first root that in all cases (somewhat unusually for infixes in Alál) replaces the root vowel entirely (though it will often show up again to force initial mid-tone, break up a cluster, or as part of a morpheme that includes partial reduplication); object plurality is included in this affix, but subject plurality is not. The persons are first, second, third, and “fourth”, where fourth is more of an indefinite third person akin to “one”, introducing new information or indicating that the argument in question isn't already in the common ground, and also to refer to less-salient third persons when there is more than one being discussed. There are also four object-less infixes that cause verbs to be considered intransitive. (This doesn't necessarily mean they have a valency of one, particularly combined with the causative and circumstantial voices, but that's a post for another time). The thirty-six total subject-object infixes are given below:
Here are some examples of them in use:
Spoiler:
Spoiler:
Spoiler:
Quantifiers indicate the general number of arguments, relative to all arguments of that type; that is: few vs. some vs. many vs. most vs. all, as well as a general plural that does not specify. With the exception of gen. plural and all, these affixes are derived from the CVC roots used for a variety of deictic constructions (including personal pronouns), defined as the proximal, medial, distal, and extra-distal roots respectively, which allows them to take diminutive and augmentative forms that vary slightly depending on that root's vowel.
A brief note on a glossing convention of mine: a lot of morphemes in Alál include, but do not consist entirely of, partial reduplication of some other element. Where this is the case, the reduplication portion at the morpheme-by-morpheme gloss level will be separated from the rest by a dot (.) and a tilde (~) in the direction of the duplicated element, combining the Leipzig rules for reduplication and fused morphemes. Additionally, “V” here refers to the vowel of the most salient root. Which root is most salient depends on where the morpheme in question is:
- If the morpheme is infixed within some root, then that root is the most salient.
- If the morpheme is a suffix and is part of or after the axis position, then the primary root will always be the most salient.
- If the morpheme is before the axis position, then the first root of the word (be it primary or secondary) is the most salient.
Spoiler:
The few-some-many-most markers (not gen. plural or all) can take what are essentially the aug./dim. oblique noun infixes (remnants of their origins as the deictic roots) to indicate “at most” and “at least”, respectively. These are -á- and -à- respectively for “some” and “many”, but -áı- and -àı- for “few” and “most” (because the proximal and extradistal both have /i/ as a root vowel, which doesn't disappear in the augmentative or diminutive oblique; see an as-of-yet unwritten post about noun cases for charts and things regarding this). And now we get some good old ininfixfixation:
Spoiler:
The second type of plural marking, numerals, is direct indication of the exact number of arguments taking part in the action. Alál allows this with single roots up to 12; beyond that, compounding is possible to produce larger or more complex numbers. The largest number it's possible to specify directly in a root is, I believe, an integer just over 9.6 * 1012 (though that's... a bit excessive). This is because Alál's number system is based on twelve integer roots (thirteen if you count zero, which I'll talk about more in a bit since it's used in negation) which can be multiplied, added, and raised to the power of each other through nominal compounding.
This, and the use of these compound numerals in plural marking, will be described later when I talk about the number system; for now, we'll stick with the basic thirteen roots 0-12. Just as with the quantifiers, the central vowel of the root is omitted unless in the augmentative or diminutive (which convey the same at least/at most meaning here), and the integers 1, 3, 4, 7, and 12 take -áı- and -àı- in those cases rather than -á- and -à-, because the roots they originate from have /i/ as a central vowel.
The numerals from 1-12 are as follows: 0 (lat), 1 (xıd), 2 (taḳ), 3 (zıṣ), 4 (ḍıṭ), 5 (ẓaḷ), 6 (tul), 7 (ṭıl), 8 (ḷuz), 9 (ḳsud), 10 (ḳaẓ), 11 (dzul), 12 (klır). These are always followed by the vowel of the most salient root. When integer plurality is used, there must also be a quantifier if an actual plural meaning is to be conveyed; typically, this is the general plural. The quantifier is infixed directly before the numeral:
Spoiler:
Spoiler:
Spoiler:
Laàtḳalaı.
la<aà-taḳ.~a>laı
<1.3-two>talk_to
I (in two pieces) talk to him.
I talk to him (but I only kind of want to).
And then there's zero, which I gloss as NEG since that's essentially what it does. There are two methods of marking negation on a verb; one is a specific negative affix which is buddies with a couple other “truth-value” affixes (used to affirm the certainty of an event's occurrence, indicate one isn't sure whether it occurred or not, or describe hypothetical situations) and the other is the use of zero as a numeral. These two methods differ greatly in scope and I'll describe how they work in more detail later. The latter method is exemplified below:
Spoiler:
Aẓîkaltaḷı kaıh.
ẓa<ı-ŕk.~ar̀-lat.~a>ḷ-ı ka<ı>h
sleep<3-all-NEG>-DUR person<NOM>
None of the people are sleeping (because there are no people).
A similarly humorous and/or tremendously unhelpful effect is achieved by putting it in the "at least" diminutive:
Ẓılàtaḷı.
ẓa<ı-la<à>t.~a>ḷ-ı.
sleep<3-NEG<DIM>>-DUR
At least zero people are sleeping.
I have absolutely no idea whether anyone is sleeping or not. Why the hell did you ask me this question?
The "at most" augmentative results in particularly strong negation, which is typically only used when one is very irritated, very commanding, or a very bad liar:
Spoiler:
Spoiler:
In the meantime, at last, we'll move on to object plurality.
Later.
In another post.
Eventually.
Last edited by kiwikami on 08 Jul 2017 18:35, edited 2 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
Re: Alál: Verbs Episode I, The Plural Menace
Wow.
Wow.
zero-number & Undefined.
...wow.
The possibilities.
walk<PST.PFT>√-1.DU
Very detailed descriptions.
Would that you had met an old colleague, Jeannie Summers (Kzhen , so that she could've illustrated your examples {RIP , PBUH & Qapla'!}
http://www.kag.org/ms/ms57/19.htm
Not exactly a linguist, but *heia...*
Wow.
zero-number & Undefined.
...wow.
The possibilities.
walk<PST.PFT>√-1.DU
Very detailed descriptions.
Would that you had met an old colleague, Jeannie Summers (Kzhen , so that she could've illustrated your examples {RIP , PBUH & Qapla'!}
http://www.kag.org/ms/ms57/19.htm
Not exactly a linguist, but *heia...*
- DesEsseintes
- mongolian
- Posts: 4331
- Joined: 31 Mar 2013 13:16
Re: Alál: Verbs Episode I, The Plural Menace
kiwikami, I am mind blown.
This is simply the most awesome conlang aesthetic I've seen so far on this board. This. Is. Beautiful.
The morphology is also extremely satisfying. I have to read this a couple more times to fully appreciate it I'm sure.
This is simply the most awesome conlang aesthetic I've seen so far on this board. This. Is. Beautiful.
The morphology is also extremely satisfying. I have to read this a couple more times to fully appreciate it I'm sure.
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Alál: Verbs Episode I, The Plural Menace
Responses to the responses:
~VERBS~
Episode II: Attack of the Clusters
Time for object plurality.
And object incorporation.
And habituals.
And a lot of vowel epenthesis.
And inininfixfixfixation.
Funnily enough this is actually a pretty short update.
Basic object plurality is indicated by the subject-object agreement marker, which is equivalent to putting the object in the general plural. Beyond this, it is expressed with a construction that looks a lot like object incorporation with a quantifier or integer marker (or both) as the object. Regardless of the actual plurality of the object, the subject-object agreement marker must indicate that the object is plural if an actual plural meaning is to be achieved (see above post on that partial-plurality, single-groups-consisting-of-multiple-members thing that happens otherwise; I'll also give more examples below). The construction consists of the partial reduplication of the first root's stem - specifically, of the first C(C)V - after any subject plurality morphemes, followed by the roots associated with either the quantifier or the numeral, minus the middle vowel. As it does not originate from a deictic root, all has its own specific object form, kù/kú, where the tone is underlying/originally low, but becomes high if the next vowel is low-tone (this shows up again with the truth value suffixes and with some forms of compounding).
For example:
Kaàḷı. I remember something. (singular object using 1.4SG aà, from the root KIḶ)
Kalàḷı. I remember some things. (general plural object using 1.4PL alà)
Kalàkıskḷı. I remember many things. (many object quantifier using 1.4PL alà, reduplication of kı-, and distal root saḷ; note ḷḷ > kḷ)
Kalàkıkùḷı. I remember everything. (all object quantifier using 1.4PL alà, reduplication of kı-, and all object form kù)
Kalàkıkûḷıtı. I can't remember everything. (all object quantifier using 1.4PL alà, reduplication of kı-, and all object form kù, negated with -ìt- suffix which causes kù to become high-tone)
Kalàkızàıṣḷı. I remember at least three things. (two object numeral using 1.4PL alà, reduplication of kı-, and two root zıṣ with diminutive infix -à-)
Kalàrkìkıklırḷı. We all remember twelve things (collectively/between us). (all subject quantifier using -ŕk.~V.r̀-; twelve object numeral using 1.4PL alà, reduplication of kı-, and twelve root klır with vowel epenthesized to break up a CCCC cluster)
Just as with subject plurality, the quantifiers and numerals may stack for the object. This is done though, essentially, nominal compounding; again, these roots are being treated as incorporated objects, which may form compounds (and take infixes, which can also take infixes... once I figure out productive locational incorporation, the maximum infixation depth will probably be five. It's three at the moment.)
Nominal compounding's specifics are a subject for another time, but suffice it to say that the "manner" compound type is what is used in this instance, meaning that the secondary and primary roots are separated by a copy of the primary root vowel; unlike in regular two-root nominal compounding, the root vowels of both roots may be deleted unless there is a diminutive or augmentative infix in either. All, incidentally, has no root vowel in this case (being not based on a deictic root). Quantifiers before numerals, just as with subjects. Thus:
all kù + seven thıl = -kùthl-
Ikûṭukùthılḷı = "You remember seven, which is all of them." (KIḶ)
all kù + seven thıl with AUG infix -á- = -kǔthaıl-
Ikûṭukǔthaılḷı = "You remember at most seven, which is all of them." (KIḶ)
most "extradistal" hız + seven thıl = -hzıthl-
Ikûṭukıhzıthılḷı = "You remember seven, which is most of them." (KIḶ)
Alûṭulahzıthıllı = "You speak to seven, which is most of them." (LAL)
Alûṭuhzathıllahzıthıllı = "You seven, which is most of you, speak to seven, which is most of them." (LAL)
Regarding Clusters, "Most Salient Roots" and Epenthesis. Note in that last example that the original word before vowel epenthesis is Alûṭuhzthllahzıthıllı, with a CCCCCC cluster (underlined - also, the red color is to show where the root LAL is still apparent,). Six-consonant clusters are twice what is permitted. YOU KNOW WHAT MUST BE DONE. Not that I don't love massive consonant clusters with all my heart, but Alál does not. The first vowel is placed before the last two consonants in the cluster; this is in the middle of the most root, so it is the vowel from that root, /i/, that is epenthesized. This still leaves a CCCC cluster, so the next vowel is put between the last two consonants of that, which is between the seven root and the most root. Since this is not between any consonants of these roots themselves, the vowel added is that of the root to which all of this is infixed, the /a/ of LAL. hzthll > hzathıll. Ta-da. *celebrates*
As said above, somewhat like with subjects, omitting the plural marking on the subject/object agreement marker while including one of these object plurality morphemes leads to a sort of partial, one-group-in-many-pieces reading; it's most used to describe employees in a company, chapters in a book, sets of dishes, and the like:
Object Incoporation and Habituals
Habituals describe something that the subject does repeatedly, over some span of time - tendencies, jobs, routines, or the like:
Object incorporation by default encodes a habitual meaning. Morphologically, it works very much like numbers: You reduplicate the C(C)V, then add the relevant root. In this case, however, the valency of the verb is decreased by one; the subject/object agreement marker does not agree for the object, instead becoming an "intransitive" infix, and any patientive argument added to the verb will be interpreted as a benefactive, not as a direct object. The incorporated root may be any noun base; nominalized verbs or the like may not be incorporated, but compound nouns or nouns with infixes are just fine, and thus you can get things like:
The <OBL> case infix appears in the above examples due to some rules regarding axis infixation in nouns (tense infixes along with direction-of-motion infixes together form a rather large group, axis markers, which can appear in both verbs and nouns and also form the bulk of the language's prepositions).
There are almost certainly a dozen things I've forgotten to mention, but I confess I'm not certain what else to talk about regarding this particular topic; I could give more examples, but that seems a bit redundant. Suggestions would be greatly welcomed. I do still have one more plural-related thing to ramble about, which are collective vs. distributive plurals, but I'll cover those in... uh... let's call it VERBS Episode III: Revenge of the DIST.
Spoiler:
~VERBS~
Episode II: Attack of the Clusters
Time for object plurality.
And object incorporation.
And habituals.
And a lot of vowel epenthesis.
And inininfixfixfixation.
Funnily enough this is actually a pretty short update.
Basic object plurality is indicated by the subject-object agreement marker, which is equivalent to putting the object in the general plural. Beyond this, it is expressed with a construction that looks a lot like object incorporation with a quantifier or integer marker (or both) as the object. Regardless of the actual plurality of the object, the subject-object agreement marker must indicate that the object is plural if an actual plural meaning is to be achieved (see above post on that partial-plurality, single-groups-consisting-of-multiple-members thing that happens otherwise; I'll also give more examples below). The construction consists of the partial reduplication of the first root's stem - specifically, of the first C(C)V - after any subject plurality morphemes, followed by the roots associated with either the quantifier or the numeral, minus the middle vowel. As it does not originate from a deictic root, all has its own specific object form, kù/kú, where the tone is underlying/originally low, but becomes high if the next vowel is low-tone (this shows up again with the truth value suffixes and with some forms of compounding).
For example:
Kaàḷı. I remember something. (singular object using 1.4SG aà, from the root KIḶ)
Kalàḷı. I remember some things. (general plural object using 1.4PL alà)
Kalàkıskḷı. I remember many things. (many object quantifier using 1.4PL alà, reduplication of kı-, and distal root saḷ; note ḷḷ > kḷ)
Kalàkıkùḷı. I remember everything. (all object quantifier using 1.4PL alà, reduplication of kı-, and all object form kù)
Kalàkıkûḷıtı. I can't remember everything. (all object quantifier using 1.4PL alà, reduplication of kı-, and all object form kù, negated with -ìt- suffix which causes kù to become high-tone)
Kalàkızàıṣḷı. I remember at least three things. (two object numeral using 1.4PL alà, reduplication of kı-, and two root zıṣ with diminutive infix -à-)
Kalàrkìkıklırḷı. We all remember twelve things (collectively/between us). (all subject quantifier using -ŕk.~V.r̀-; twelve object numeral using 1.4PL alà, reduplication of kı-, and twelve root klır with vowel epenthesized to break up a CCCC cluster)
Just as with subject plurality, the quantifiers and numerals may stack for the object. This is done though, essentially, nominal compounding; again, these roots are being treated as incorporated objects, which may form compounds (and take infixes, which can also take infixes... once I figure out productive locational incorporation, the maximum infixation depth will probably be five. It's three at the moment.)
Nominal compounding's specifics are a subject for another time, but suffice it to say that the "manner" compound type is what is used in this instance, meaning that the secondary and primary roots are separated by a copy of the primary root vowel; unlike in regular two-root nominal compounding, the root vowels of both roots may be deleted unless there is a diminutive or augmentative infix in either. All, incidentally, has no root vowel in this case (being not based on a deictic root). Quantifiers before numerals, just as with subjects. Thus:
all kù + seven thıl = -kùthl-
Ikûṭukùthılḷı = "You remember seven, which is all of them." (KIḶ)
all kù + seven thıl with AUG infix -á- = -kǔthaıl-
Ikûṭukǔthaılḷı = "You remember at most seven, which is all of them." (KIḶ)
most "extradistal" hız + seven thıl = -hzıthl-
Ikûṭukıhzıthılḷı = "You remember seven, which is most of them." (KIḶ)
Alûṭulahzıthıllı = "You speak to seven, which is most of them." (LAL)
Alûṭuhzathıllahzıthıllı = "You seven, which is most of you, speak to seven, which is most of them." (LAL)
Regarding Clusters, "Most Salient Roots" and Epenthesis. Note in that last example that the original word before vowel epenthesis is Alûṭuhzthllahzıthıllı, with a CCCCCC cluster (underlined - also, the red color is to show where the root LAL is still apparent,). Six-consonant clusters are twice what is permitted. YOU KNOW WHAT MUST BE DONE. Not that I don't love massive consonant clusters with all my heart, but Alál does not. The first vowel is placed before the last two consonants in the cluster; this is in the middle of the most root, so it is the vowel from that root, /i/, that is epenthesized. This still leaves a CCCC cluster, so the next vowel is put between the last two consonants of that, which is between the seven root and the most root. Since this is not between any consonants of these roots themselves, the vowel added is that of the root to which all of this is infixed, the /a/ of LAL. hzthll > hzathıll. Ta-da. *celebrates*
As said above, somewhat like with subjects, omitting the plural marking on the subject/object agreement marker while including one of these object plurality morphemes leads to a sort of partial, one-group-in-many-pieces reading; it's most used to describe employees in a company, chapters in a book, sets of dishes, and the like:
Spoiler:
Object Incoporation and Habituals
Habituals describe something that the subject does repeatedly, over some span of time - tendencies, jobs, routines, or the like:
Spoiler:
Spoiler:
There are almost certainly a dozen things I've forgotten to mention, but I confess I'm not certain what else to talk about regarding this particular topic; I could give more examples, but that seems a bit redundant. Suggestions would be greatly welcomed. I do still have one more plural-related thing to ramble about, which are collective vs. distributive plurals, but I'll cover those in... uh... let's call it VERBS Episode III: Revenge of the DIST.
Last edited by kiwikami on 08 Jul 2017 18:40, edited 1 time in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Alál: Verbs Episode II, Attack of the Clusters
~VERBS~
Episode III: Revenge of the DIST
One more quick thing and then I swear I'm done with plurality. I promise.
There is a distinction between the collective and distributive plural, where the former indicates all subjects or objects are acting/being acted upon together and the latter that they are doing so separately. The collective is the default; the distributive is marked through internal (infixed - no surprise there, really) reduplication of the quantifier (collective "many [subject]s" = s(a)ḷ, distributive "many [subject]s" = ssḷḷ > ssakḷ) This comes from an old augmentative construction that'll show up again in the axis markers to mark large distances, and some now-unproductive noun derivations (has/huıs/uhùs fire.OBL/AGT/PAT > uhháss/huhııss/huhuss wildfire/conflagration.OBL/AGT/PAT). Distributive plural objects without a quantifier, which use only the subject/object agreement infix to indicate their plurality, have that infix reduplicated in the same way:
Kzılìẓrakzazṣẓılàa.
kza<ılì-ẓar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL-some~three_obj>
Collective Subject, Collective Object: Some of them hunted and caught three things together, all at once.
Kzıılìlıẓrazṣẓılàa.
kza<ı<ılì>lì-ẓar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL<DIST>-some~three_obj>
Collective Subject, Distributive Object: Some of them hunted and caught three things together, one after the other.
This essentially describes three separate instances of hunting and catching. Kzıízrẓılàa Kzıízrẓılàa Kzıízrẓılàa,
"some of them hunted and caught something" repeated thrice, would be for the most part semantically equivalent.
Kzılìdẓarrakzazṣẓılàa.
kza<ılì-ẓ<ẓar̀>ar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL-some<DIST>~three_obj>
Distributive Subject, Collective Object: Some of them hunted and caught three things each, all at once.
Note the vowel epenthesis, marked in red, to break up a cluster; the /a/ marked in green is re-added for the same reason, though this is an echo of the /a/ from ẓar̀ rather the one from kzaẓ. This green /a/ was, in fact, deleted, as the middle vowels of the quantifiers tend to be, but returns to fix the cluster issue; we know that it wasn't just not deleted, because if that were the case, we would've seen ar̀ > à. But vowel epenthesis happens after the floating "tones" look for a home, so this is in fact a re-appearance of the deleted vowel.
Kzıılìlıdẓarrakzazṣẓılàa.
kza<ı<ılì>lì-ẓ<ẓar̀>ar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL<DIST>-some<DIST>~three_obj>
Distributive Subject, Distributive Object: Some of them hunted and caught three things each, one after the other.
Three separate "hunting and catching" events are occurring for each subject; were we to specify the number of the subject, saying for example "two of them" instead of "some of them", there'd be a total of six separate events.
There are a handful of sound change rules that often come into play (the affricate+affricate=stop+affricate rule and vowel epenthesis, most notably) during this which also show up in later instances of internal reduplication, particularly with the frequentative aspect and some thing that happen to the direction-of-motion and tense markers. The only one that isn't already listed as a sound change in the phonology is what happens to non-M vowels: If the reduplicated morpheme has non-M vowels, only the infixed part will keep those vowels, while the outer part will not; they will reduce to their M forms. Thus, ılì > ıılìlı, íḳ > ıíḳḳ, kxù > kkxùxu, etc.
Again, I'm certain there's more to say, but I'm not sure what. Hm. Well. As previously said, one can indicate multiple subjects/objects of different persons by stacking up to two of the sub./obj. agreement markers. Alongside the distributive plural, this can lead to some fantastically ridiculous-looking words. I shall share one of my favorites.
Lıılìlıılǐlırlıılıılıaı.
la<ı<ılì>lì-ı<ılì>lì-ŕ<ŕ>>l-ı<ıl-ıı>l~-ı<ı>-a
speak<3.3PL<DIST>-3.3PL<DIST>-PL<DIST>>-PST<only-AUG>-DUR<FREQ>-VOL.ACT
A long time ago, but never since, they went around one after another talking to themselves.
So that's that.
Episode III: Revenge of the DIST
One more quick thing and then I swear I'm done with plurality. I promise.
There is a distinction between the collective and distributive plural, where the former indicates all subjects or objects are acting/being acted upon together and the latter that they are doing so separately. The collective is the default; the distributive is marked through internal (infixed - no surprise there, really) reduplication of the quantifier (collective "many [subject]s" = s(a)ḷ, distributive "many [subject]s" = ssḷḷ > ssakḷ) This comes from an old augmentative construction that'll show up again in the axis markers to mark large distances, and some now-unproductive noun derivations (has/huıs/uhùs fire.OBL/AGT/PAT > uhháss/huhııss/huhuss wildfire/conflagration.OBL/AGT/PAT). Distributive plural objects without a quantifier, which use only the subject/object agreement infix to indicate their plurality, have that infix reduplicated in the same way:
Kzılìẓrakzazṣẓılàa.
kza<ılì-ẓar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL-some~three_obj>
Collective Subject, Collective Object: Some of them hunted and caught three things together, all at once.
Kzıılìlıẓrazṣẓılàa.
kza<ı<ılì>lì-ẓar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL<DIST>-some~three_obj>
Collective Subject, Distributive Object: Some of them hunted and caught three things together, one after the other.
This essentially describes three separate instances of hunting and catching. Kzıízrẓılàa Kzıízrẓılàa Kzıízrẓılàa,
"some of them hunted and caught something" repeated thrice, would be for the most part semantically equivalent.
Kzılìdẓarrakzazṣẓılàa.
kza<ılì-ẓ<ẓar̀>ar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL-some<DIST>~three_obj>
Distributive Subject, Collective Object: Some of them hunted and caught three things each, all at once.
Note the vowel epenthesis, marked in red, to break up a cluster; the /a/ marked in green is re-added for the same reason, though this is an echo of the /a/ from ẓar̀ rather the one from kzaẓ. This green /a/ was, in fact, deleted, as the middle vowels of the quantifiers tend to be, but returns to fix the cluster issue; we know that it wasn't just not deleted, because if that were the case, we would've seen ar̀ > à. But vowel epenthesis happens after the floating "tones" look for a home, so this is in fact a re-appearance of the deleted vowel.
Kzıılìlıdẓarrakzazṣẓılàa.
kza<ı<ılì>lì-ẓ<ẓar̀>ar̀~kza.zıṣ>ẓılàa
hunted_and_caught<3.3PL<DIST>-some<DIST>~three_obj>
Distributive Subject, Distributive Object: Some of them hunted and caught three things each, one after the other.
Three separate "hunting and catching" events are occurring for each subject; were we to specify the number of the subject, saying for example "two of them" instead of "some of them", there'd be a total of six separate events.
There are a handful of sound change rules that often come into play (the affricate+affricate=stop+affricate rule and vowel epenthesis, most notably) during this which also show up in later instances of internal reduplication, particularly with the frequentative aspect and some thing that happen to the direction-of-motion and tense markers. The only one that isn't already listed as a sound change in the phonology is what happens to non-M vowels: If the reduplicated morpheme has non-M vowels, only the infixed part will keep those vowels, while the outer part will not; they will reduce to their M forms. Thus, ılì > ıılìlı, íḳ > ıíḳḳ, kxù > kkxùxu, etc.
Again, I'm certain there's more to say, but I'm not sure what. Hm. Well. As previously said, one can indicate multiple subjects/objects of different persons by stacking up to two of the sub./obj. agreement markers. Alongside the distributive plural, this can lead to some fantastically ridiculous-looking words. I shall share one of my favorites.
Lıılìlıılǐlırlıılıılıaı.
la<ı<ılì>lì-ı<ılì>lì-ŕ<ŕ>>l-ı<ıl-ıı>l~-ı<ı>-a
speak<3.3PL<DIST>-3.3PL<DIST>-PL<DIST>>-PST<only-AUG>-DUR<FREQ>-VOL.ACT
A long time ago, but never since, they went around one after another talking to themselves.
So that's that.
Last edited by kiwikami on 13 Apr 2018 19:33, edited 1 time in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Alál: Negation and Other Such Nonsense
∀x∀y[(Kiwikami(x) & Alál(y)) → ~working.on(x,y)]
An Interlude In Which I Pretend I Remember What I Learned In Semantics Class
----------
Having found (for the second time after losing it) my old notebook on this conlang, I figure it's time I got back to recording everything in it so that it can no longer be lost to the void. The bad news is that it's been a while and I only barely remember how anything except basic verbal morphology works. The good news is that I left off in the middle of a bunch of verbal morphology, so I can pick up there. But first... negation. It needs more explanation, because the more I read my notes the more I realize what a mess the system is/was, and it needs to be written out clearly so that I can inevitably completely change it.
This is an attempt to do this thing. The issue: I wrote these notes back when I was taking undergrad semantics. I haven't touched semantics in three years. So I do apologize if nothing that I say here makes any sense. The point of getting it on the computer is so that it is easier to revise, anyhow, and I most definitely welcome suggestions for revision.
----------
Of Death, Water, and Things That Aren't True
I mentioned a few posts back that there are two methods of marking negation: one is indicating a subject of number zero, and one is a truth-value affix appearing later. The first one I described to some extent then, but here I'll get to the second of these. This affix is -ìt-, and it goes immediately after the first verb root. If the preceding and following vowels are both low-marked, it becomes -ít-. It is one of a small handful of these affixes, including an affirmational one used predominantly for focus (especially contrastive focus), an undefined parallel to the undefined numeral shown earlier, and a hypothetical that I'll get to later. In describing the odd behavior of -ìt- in relation to verbal compounding, it is worth noting here that the affirmational/hypothetical affixes do not behave the same way as the negative/undefined in terms of scope; for the latter, as described shortly, the scope is restricted to the first verb, while for the former it is over the entire verb.
Here are some examples of how it works, contrasted with the lat.~V zero-negation. The truth-value negation has, in several ways, a different scope than the zero-negation. This is a two-part issue; in non-compounded words, there are semantic shenanigans to consider which I'll discuss in a bit. In compound words, the distinction is that while zero-negation negates the primary verb (or, arguably, the entire statement), truth-value negation negates the first verb. In compounds, these are not the same thing, and the result is that any compound verb containing truth-value negation is not imposing falseness on the whole statement; it's not actually negative. The statement as a whole is a positive one.
Tzazìtṣıḳaàl.
tza<a>z-ìt-ṣıḳaàl
<1>be.in.water-NEG-be.dying
I am dying of thirst / dying due to not being in [contact with] water.
Tzaltazṣıḳaàl.
tza<a-lat.~a>z-ṣıḳaàl
<1-NEG>be.in.water-be.dying
I am not drowning / not dying due to being in [contact with] water.
So, in compounds (and only in compounds, mind you), truth-value negation isn't really negation. The different behaviors of this and the affirmational/hypothetical affixes (which do apply to the entire statement) is a "relatively recent" "historical" development in the language. It is partially a result of the increased presence of zero and undefined numeral markers used in negation and denial of presupposition, but more importantly due to the decreased productivity of the negative truth-value affix in verbal compounding (and of verbal compounding in general) due to semantic drift. The affirmational/hypothetical/undefined affixes were not used in compounds nearly to the extent that the negative was. The affirmational/hypothetical, due to the lack of any other constructions that gave these qualities to the entire verb, thus grew to do so (though there are still a small handful of compound verbs that use affirmational forms).
At the same time, negative-using compounds, through semantic drift, became less easily broken into their constituent parts, and the -ìt- morpheme grew less productive; with an alternative form of negation possible through the zero numeral, and with its presence growing increasingly frozen into the compound verb forms, -ìt- never generalized. Partially through analogy, and partially because it also had a numeral-using alternative, the undefined affix did not, either. You can, of course, negate compound verb -ìt- constructions - with the zero numeral:
Tzaltazìtṣıḳaàl.
tza<a-lat.~a>z-ìt-ṣıḳaàl
<1-NEG>be.in.water-NEG-be.dying
I am not dying of thirst.
In non-compound verbs with no quantifiers attached to the person markers, truth-value negation works, as always, on the first verb - which in this case happens also to be the primary verb, and thus the entire statement is negated. Here the result is actually negative, but it has a different scope than zero-negation. The distinction is as follows:
Rasìtaı.
ra<a>s-ìt-aı
<1>go-NEG-be.going
I am not-going.
∃x[be.me(x) & ~going(x)]
Raltasaı.
ra<a-lat.~a>saı
<1-NEG>be.going
I am not going.
~∃x[be.me(x) & going(x)]
Zero-negation is preferred here, with truth-value negation used when one wishes to imply that one is doing something other than the verb in question or that one is pointedly avoiding the action. Truth-value negation in quantifier-less non-compounds is rare, though it does show up if the verb uses a subject-modifying numeral (as one is blocked from also using zero-negation where there is a non-zero numeral) in constructions like Arùtḳasìtaı "You two are not going."
There is also a small group of non-compound verbs with lexified truth-value negation; these are identifiable through the use of a slightly different volitional suffix. In the active voice, this is either á as in Rasìtıá, "I am staying here"/"I am [pointedly] not-going" and Ḳalîtaá "I survived/"I [pointedly] did not die", or ŕx.~Vr̀ as in Dzaâıtîxa "I am completely uncomprehending of it"/"I am [reluctantly] not-understanding it" and Xıxǐtıxà "it has been made harmless"/"it [reluctantly] is not-dangerous".
In non-compound verbs with a quantifier attached to the person marker, the two forms are somewhat more evenly distributed as the truth-value X is not-going form has more useful implications. As seen with be.going above, the scope of truth-value negation does not include the subject. Rasìtaı would be false if the speaker did not exist, while Raltasaı would not (we're ignoring presupposition here for a moment, and however the heck pronouns are supposed to work, to prove a point). If one takes the quantifier all (ŕk.~Vr̀), then any statement made using zero-negation is a claim that it is not the case that all X are doing something; any statement with truth-value negation is one that it is the case that all X are not doing something. The result is essentially translatable as the distinction between "Not all of them are going" and "None of them are going", as seen below:
Arîkasìtaı
ra<ı-ŕk.~a.r̀>s-ìt-aı
go<3-all>-NEG-be.going
They are all not-going.
It is the case that all of them are not going. ("None of them are going.")
Arîkaltasaı
ra<ı-ŕk.~a.r̀-lat.~a>saı
go<3-all-NEG>
They are not all going.
It is not the case that all of them are going. ("They aren't all going [but some might be].")
But wait, Kiwi, I hear you (probably not) ask; what happens when you have a compound verb that also has a quantifier? The answer is that I didn't write down an answer in this mess of a notebook, or if I did, it's too far buried under multiple layers of crossings-out in sharpie, decisions to actually use previously-crossed-out things notated with a THIS IS FINE and an arrow in the margins, and decisions to cross the THIS IS FINE note out and therefore render the actual information again crossed-out... followed by further rescinding of that decision, all while the rest of the space on the page is filled with a lovely chart of verbal affixes that doesn't match another chart purportedly providing the exact same affixes three pages later.
But given what I've decided here regarding the "historical" basis of the two negation methods, I think it's safe to say that in such a case, it behaves just as it does in non-quantifier compound verbs. Whether it is or isn't the case that some individuals may be performing the action (though all are not) would be ambiguous. I'm fine with that. Ambiguity is good fun.
Atzîkaltazṣıḳaàl.
tza<ı-ŕk.~a.r̀-lat.~a>z-ṣıḳaàl
<3-all-NEG>be.in.water-NEG-be.dying
It is not the case that all of them are dying of thirst.
or None of them are dying of thirst.
or Some but not all of them are dying of thirst.
Next up, I'll talk about the other truth-value affixes in VERBS Episode IV: A New HYP.
An Interlude In Which I Pretend I Remember What I Learned In Semantics Class
----------
Having found (for the second time after losing it) my old notebook on this conlang, I figure it's time I got back to recording everything in it so that it can no longer be lost to the void. The bad news is that it's been a while and I only barely remember how anything except basic verbal morphology works. The good news is that I left off in the middle of a bunch of verbal morphology, so I can pick up there. But first... negation. It needs more explanation, because the more I read my notes the more I realize what a mess the system is/was, and it needs to be written out clearly so that I can inevitably completely change it.
This is an attempt to do this thing. The issue: I wrote these notes back when I was taking undergrad semantics. I haven't touched semantics in three years. So I do apologize if nothing that I say here makes any sense. The point of getting it on the computer is so that it is easier to revise, anyhow, and I most definitely welcome suggestions for revision.
----------
Of Death, Water, and Things That Aren't True
I mentioned a few posts back that there are two methods of marking negation: one is indicating a subject of number zero, and one is a truth-value affix appearing later. The first one I described to some extent then, but here I'll get to the second of these. This affix is -ìt-, and it goes immediately after the first verb root. If the preceding and following vowels are both low-marked, it becomes -ít-. It is one of a small handful of these affixes, including an affirmational one used predominantly for focus (especially contrastive focus), an undefined parallel to the undefined numeral shown earlier, and a hypothetical that I'll get to later. In describing the odd behavior of -ìt- in relation to verbal compounding, it is worth noting here that the affirmational/hypothetical affixes do not behave the same way as the negative/undefined in terms of scope; for the latter, as described shortly, the scope is restricted to the first verb, while for the former it is over the entire verb.
Here are some examples of how it works, contrasted with the lat.~V zero-negation. The truth-value negation has, in several ways, a different scope than the zero-negation. This is a two-part issue; in non-compounded words, there are semantic shenanigans to consider which I'll discuss in a bit. In compound words, the distinction is that while zero-negation negates the primary verb (or, arguably, the entire statement), truth-value negation negates the first verb. In compounds, these are not the same thing, and the result is that any compound verb containing truth-value negation is not imposing falseness on the whole statement; it's not actually negative. The statement as a whole is a positive one.
Tzazìtṣıḳaàl.
tza<a>z-ìt-ṣıḳaàl
<1>be.in.water-NEG-be.dying
I am dying of thirst / dying due to not being in [contact with] water.
Tzaltazṣıḳaàl.
tza<a-lat.~a>z-ṣıḳaàl
<1-NEG>be.in.water-be.dying
I am not drowning / not dying due to being in [contact with] water.
So, in compounds (and only in compounds, mind you), truth-value negation isn't really negation. The different behaviors of this and the affirmational/hypothetical affixes (which do apply to the entire statement) is a "relatively recent" "historical" development in the language. It is partially a result of the increased presence of zero and undefined numeral markers used in negation and denial of presupposition, but more importantly due to the decreased productivity of the negative truth-value affix in verbal compounding (and of verbal compounding in general) due to semantic drift. The affirmational/hypothetical/undefined affixes were not used in compounds nearly to the extent that the negative was. The affirmational/hypothetical, due to the lack of any other constructions that gave these qualities to the entire verb, thus grew to do so (though there are still a small handful of compound verbs that use affirmational forms).
At the same time, negative-using compounds, through semantic drift, became less easily broken into their constituent parts, and the -ìt- morpheme grew less productive; with an alternative form of negation possible through the zero numeral, and with its presence growing increasingly frozen into the compound verb forms, -ìt- never generalized. Partially through analogy, and partially because it also had a numeral-using alternative, the undefined affix did not, either. You can, of course, negate compound verb -ìt- constructions - with the zero numeral:
Tzaltazìtṣıḳaàl.
tza<a-lat.~a>z-ìt-ṣıḳaàl
<1-NEG>be.in.water-NEG-be.dying
I am not dying of thirst.
In non-compound verbs with no quantifiers attached to the person markers, truth-value negation works, as always, on the first verb - which in this case happens also to be the primary verb, and thus the entire statement is negated. Here the result is actually negative, but it has a different scope than zero-negation. The distinction is as follows:
Rasìtaı.
ra<a>s-ìt-aı
<1>go-NEG-be.going
I am not-going.
∃x[be.me(x) & ~going(x)]
Raltasaı.
ra<a-lat.~a>saı
<1-NEG>be.going
I am not going.
~∃x[be.me(x) & going(x)]
Zero-negation is preferred here, with truth-value negation used when one wishes to imply that one is doing something other than the verb in question or that one is pointedly avoiding the action. Truth-value negation in quantifier-less non-compounds is rare, though it does show up if the verb uses a subject-modifying numeral (as one is blocked from also using zero-negation where there is a non-zero numeral) in constructions like Arùtḳasìtaı "You two are not going."
There is also a small group of non-compound verbs with lexified truth-value negation; these are identifiable through the use of a slightly different volitional suffix. In the active voice, this is either á as in Rasìtıá, "I am staying here"/"I am [pointedly] not-going" and Ḳalîtaá "I survived/"I [pointedly] did not die", or ŕx.~Vr̀ as in Dzaâıtîxa "I am completely uncomprehending of it"/"I am [reluctantly] not-understanding it" and Xıxǐtıxà "it has been made harmless"/"it [reluctantly] is not-dangerous".
In non-compound verbs with a quantifier attached to the person marker, the two forms are somewhat more evenly distributed as the truth-value X is not-going form has more useful implications. As seen with be.going above, the scope of truth-value negation does not include the subject. Rasìtaı would be false if the speaker did not exist, while Raltasaı would not (we're ignoring presupposition here for a moment, and however the heck pronouns are supposed to work, to prove a point). If one takes the quantifier all (ŕk.~Vr̀), then any statement made using zero-negation is a claim that it is not the case that all X are doing something; any statement with truth-value negation is one that it is the case that all X are not doing something. The result is essentially translatable as the distinction between "Not all of them are going" and "None of them are going", as seen below:
Arîkasìtaı
ra<ı-ŕk.~a.r̀>s-ìt-aı
go<3-all>-NEG-be.going
They are all not-going.
It is the case that all of them are not going. ("None of them are going.")
Arîkaltasaı
ra<ı-ŕk.~a.r̀-lat.~a>saı
go<3-all-NEG>
They are not all going.
It is not the case that all of them are going. ("They aren't all going [but some might be].")
But wait, Kiwi, I hear you (probably not) ask; what happens when you have a compound verb that also has a quantifier? The answer is that I didn't write down an answer in this mess of a notebook, or if I did, it's too far buried under multiple layers of crossings-out in sharpie, decisions to actually use previously-crossed-out things notated with a THIS IS FINE and an arrow in the margins, and decisions to cross the THIS IS FINE note out and therefore render the actual information again crossed-out... followed by further rescinding of that decision, all while the rest of the space on the page is filled with a lovely chart of verbal affixes that doesn't match another chart purportedly providing the exact same affixes three pages later.
But given what I've decided here regarding the "historical" basis of the two negation methods, I think it's safe to say that in such a case, it behaves just as it does in non-quantifier compound verbs. Whether it is or isn't the case that some individuals may be performing the action (though all are not) would be ambiguous. I'm fine with that. Ambiguity is good fun.
Atzîkaltazṣıḳaàl.
tza<ı-ŕk.~a.r̀-lat.~a>z-ṣıḳaàl
<3-all-NEG>be.in.water-NEG-be.dying
It is not the case that all of them are dying of thirst.
or None of them are dying of thirst.
or Some but not all of them are dying of thirst.
Next up, I'll talk about the other truth-value affixes in VERBS Episode IV: A New HYP.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Alál: Phonology, Part 2
Phonology, Part 2: Vowels
This is long overdue and very abbreviated; I'm procrastinating.
Vowels may be marked as H, M, or L; this isn't a tone distinction (though it used to be, thus the letter choice), but a combination of height and tendency to attract stress, where M could be considered "reduced" and unstressed lax forms. H is marked with an acute accent, L with a grave accent, and M is unmarked. Corresponding SRs for each vowel in each form are roughly as follows:
This could be considered a nine-vowel system. I don't see it as such, for several reasons. First, roots only ever contain M vowels, while a wide variety of morphological processes lead to vowels changing form; thus, it seems useful to consider HML as a separate value added on top of the three base vowels. Secondly, the H, M, and L groups do not form exclusive natural classes. The only commonality between [u] and [æ], for example, is that both are H forms; they differ in height, backness, roundedness, and ATR to boot.
The first vowel of a word must always be underlyingly M. If this is naturally the case, the M is realized as if it were L (M>L/#C(C)_), except when it is the first in a vowel sequence (as in tkıìr):
tzaz [tʃɐʂ] 'water.OBL' > [tʃɑʂ]
Tırı ['t̪ɛ.ɾɛ] 'it is dirty' > ['t̪e.ɾɛ]
tkıìr [t̪kjɛɾ] 'extended family.NOM'
If the vowel is instead L or H, then the root vowel must be epenthesized word-initially. This occurs after M>L/#C(C)_, so the epenthesized vowel remains M:
tzùaz 'water.PAT' > atzùaz [ɐ'tʃoʕ̞ʂ], where the root vowel is a from √TZAZ 'water'
Tùrı ['t̪o.ɾɛ] 'you are dirty' > Utùrı [ɔ't̪o.ɾɛ], where the root vowel is u from √TUR 'soil'
tkár [t̪kæɾ] 'extended family.OBL' > ıtkár [ɛt̪'kæɾ], where the root vowel is ı from √TKIR 'nest'
There are no real diphthongs, but pairs or longer strings of vowels do appear. HH sequences are separate syllables, as are LL sequences; typically, the first will be stressed while the second will not. Thus, Lılılíá [ɬe.ɮɛ'ɬi.æ] 'he is talkative' and Rıssùà [ɾes'so.ɑ] 'he curls in on himself (from pain)'. Such sequences are dispreferred, and there are a variety of morpheme-specific processes that avoid HH and LL sequences in contexts where they tend to appear. An examples is the negation marker ìt, which, when in an LL sequence, forces the first L in that sequence (be it ìt or some preceding vowel) to become H. Thus, *Ilùlìtıá [ɛ'ɬo.ɮe.d̪jæ] > Ilûlıtıá [ɛ.ɮu'ɬe.d̪jæ] 'you are pointedly not speaking'. Note the shift in stress here as well, since it will appear on the first L (stress is described in more detail below).
As seen in the previous examples, there's an orthographical convention by which HL sequences are represented with a circumflex on H, while L is unmarked; the same occurs with a caron on L in LH sequences (Ulǔhı [ɔ'ɬo.ʝi] 'you are alive'). This is a holdover from when HML was purely a tonal distinction, and there was some downstepping and other shenanigans involved. When I decided out-of-world to switch over to the height-and-stress system that it currently is, the diacritics were kept for clarity's sake during the intermediary stage where, rather than a three-way height distinction, there was a two-way distinction; originally, M and L both had L pronunciations, and differed only in their ability to take stress, but there were special rules for HL and LH sequences. These diacritics are no longer needed to mark these rules, but they've become a bit of an orthographic fossil. If there is an HL or LH sequence of the same height at the end of a word, the first vowel takes the diacritic, and the second is unwritten.
In HM, LM, MH, or ML vowel pairs, the M becomes an approximant ([w], [j], or [ʕ̞]). In MM pairs, the second M becomes an approximant, save when said pair is word-initial (in which case, the first M become an approximant). This was seen above with atzùaz [ɐ'tʃoʕ̞ʂ] 'water.PAT', and again in Ilûlıtıá [ɛ.ɮu'ɬe.d̪jæ] 'you are pointedly not speaking'; an example with [w] is ulíuh [ɔ'ɬiwç] 'small animal.AGT'. There are morphophonological special cases, particularly with the person markers when both vowels in such a pair are identical (in terms of the underlying ı-u-a); for instance, the 1st person subject, 3rd singular object infix -aà- is pronounced [ɑː], while the 3rd subject, 3rd singular object -ıí- is [iː], rather than the expected [ʕ̞ɑ] and [ji]. There are only two other such cases, both combinations of the conclusive aspect -à- and a volitive -a- or -á-, together -àa- [ɑː] or -ǎ- [ʕ̞æ]. We thus see [ɑː] twice in Ḳaàlàa ['ᵑkʼɑː.ɮɑː] 'I just killed it', where the àa sequence is the conclusive+volitive, but only once in Raàstàa [ʼɾɑːs.tɑʕ̞] 'I just took it down from up there' where the sequence is a combination of the conclusive and a locational infix.
Each word has one primary stress, determined as follows:
This is long overdue and very abbreviated; I'm procrastinating.
Vowels may be marked as H, M, or L; this isn't a tone distinction (though it used to be, thus the letter choice), but a combination of height and tendency to attract stress, where M could be considered "reduced" and unstressed lax forms. H is marked with an acute accent, L with a grave accent, and M is unmarked. Corresponding SRs for each vowel in each form are roughly as follows:
This could be considered a nine-vowel system. I don't see it as such, for several reasons. First, roots only ever contain M vowels, while a wide variety of morphological processes lead to vowels changing form; thus, it seems useful to consider HML as a separate value added on top of the three base vowels. Secondly, the H, M, and L groups do not form exclusive natural classes. The only commonality between [u] and [æ], for example, is that both are H forms; they differ in height, backness, roundedness, and ATR to boot.
The first vowel of a word must always be underlyingly M. If this is naturally the case, the M is realized as if it were L (M>L/#C(C)_), except when it is the first in a vowel sequence (as in tkıìr):
tzaz [tʃɐʂ] 'water.OBL' > [tʃɑʂ]
Tırı ['t̪ɛ.ɾɛ] 'it is dirty' > ['t̪e.ɾɛ]
tkıìr [t̪kjɛɾ] 'extended family.NOM'
If the vowel is instead L or H, then the root vowel must be epenthesized word-initially. This occurs after M>L/#C(C)_, so the epenthesized vowel remains M:
tzùaz 'water.PAT' > atzùaz [ɐ'tʃoʕ̞ʂ], where the root vowel is a from √TZAZ 'water'
Tùrı ['t̪o.ɾɛ] 'you are dirty' > Utùrı [ɔ't̪o.ɾɛ], where the root vowel is u from √TUR 'soil'
tkár [t̪kæɾ] 'extended family.OBL' > ıtkár [ɛt̪'kæɾ], where the root vowel is ı from √TKIR 'nest'
There are no real diphthongs, but pairs or longer strings of vowels do appear. HH sequences are separate syllables, as are LL sequences; typically, the first will be stressed while the second will not. Thus, Lılılíá [ɬe.ɮɛ'ɬi.æ] 'he is talkative' and Rıssùà [ɾes'so.ɑ] 'he curls in on himself (from pain)'. Such sequences are dispreferred, and there are a variety of morpheme-specific processes that avoid HH and LL sequences in contexts where they tend to appear. An examples is the negation marker ìt, which, when in an LL sequence, forces the first L in that sequence (be it ìt or some preceding vowel) to become H. Thus, *Ilùlìtıá [ɛ'ɬo.ɮe.d̪jæ] > Ilûlıtıá [ɛ.ɮu'ɬe.d̪jæ] 'you are pointedly not speaking'. Note the shift in stress here as well, since it will appear on the first L (stress is described in more detail below).
As seen in the previous examples, there's an orthographical convention by which HL sequences are represented with a circumflex on H, while L is unmarked; the same occurs with a caron on L in LH sequences (Ulǔhı [ɔ'ɬo.ʝi] 'you are alive'). This is a holdover from when HML was purely a tonal distinction, and there was some downstepping and other shenanigans involved. When I decided out-of-world to switch over to the height-and-stress system that it currently is, the diacritics were kept for clarity's sake during the intermediary stage where, rather than a three-way height distinction, there was a two-way distinction; originally, M and L both had L pronunciations, and differed only in their ability to take stress, but there were special rules for HL and LH sequences. These diacritics are no longer needed to mark these rules, but they've become a bit of an orthographic fossil. If there is an HL or LH sequence of the same height at the end of a word, the first vowel takes the diacritic, and the second is unwritten.
In HM, LM, MH, or ML vowel pairs, the M becomes an approximant ([w], [j], or [ʕ̞]). In MM pairs, the second M becomes an approximant, save when said pair is word-initial (in which case, the first M become an approximant). This was seen above with atzùaz [ɐ'tʃoʕ̞ʂ] 'water.PAT', and again in Ilûlıtıá [ɛ.ɮu'ɬe.d̪jæ] 'you are pointedly not speaking'; an example with [w] is ulíuh [ɔ'ɬiwç] 'small animal.AGT'. There are morphophonological special cases, particularly with the person markers when both vowels in such a pair are identical (in terms of the underlying ı-u-a); for instance, the 1st person subject, 3rd singular object infix -aà- is pronounced [ɑː], while the 3rd subject, 3rd singular object -ıí- is [iː], rather than the expected [ʕ̞ɑ] and [ji]. There are only two other such cases, both combinations of the conclusive aspect -à- and a volitive -a- or -á-, together -àa- [ɑː] or -ǎ- [ʕ̞æ]. We thus see [ɑː] twice in Ḳaàlàa ['ᵑkʼɑː.ɮɑː] 'I just killed it', where the àa sequence is the conclusive+volitive, but only once in Raàstàa [ʼɾɑːs.tɑʕ̞] 'I just took it down from up there' where the sequence is a combination of the conclusive and a locational infix.
Each word has one primary stress, determined as follows:
- If there is at least one L that is not the final vowel (and not in a final conclusive+volitive ǎ), choose the first L. (Uhìhusí [ɔ'çi.ʝɔˌsi] 'The situation is FUBAR (idiomatic)')
- Else, if there is at least one H, choose the first H. (alál [ɐ'ɬæɬ] 'language.AUG')
- Else, if there is an L as a final vowel (or in a final conclusive+volitive ǎ), choose that L. (ıkùs [ɛ'kos] 'child.PAT'')
- Else, choose the first vowel. (hıs [çes] 'fire.AGT')
- Ahùkaxıá [ɐ'ço.gɐxˌjæ] 'you are flying low(er than expected)' - Secondary stress shows up on the final á.
- Ahùkxǎhıá [ɐ'ço.kxɑˌçi.æ] 'you are recovering from stalling and plummeting downwards' - The ǎ is L, but does not take secondary stress because it is adjacent to the stressed u. The following ı is H, and takes secondary stress.
- Ahùkxaızîrxa [ɐ'ço.kxɐj.ʂiɾˌxɑ] 'you are weak/sickly/in danger of being left for dead (idiomatic)' - The î is H, and could take secondary stress, but it is followed by a, which is L and therefore preferred.
- Ahǔkatǎ [ɐ'ço.gæ.d̪ʕ̞æ] 'you ascend (while flying)' - the central a is H, but does not take secondary stress because it is adjacent to the stressed u. The final ǎ is an LH sequence, but as a conclusive+volitive pair in a word that has another non-M vowel, it takes no stress.
Last edited by kiwikami on 29 Apr 2018 21:19, edited 3 times in total.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |
- Creyeditor
- MVP
- Posts: 5091
- Joined: 14 Aug 2012 19:32
Re: Alál: Vowels Exist
OMG. I like these vowels so much, my nose starts bleeding.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 2 3 4 4
Ook & Omlűt & Nautli languages & Sperenjas
Papuan languages, Morphophonology, Lexical Semantics
"Thoughts are free."
Produce, Analyze, Manipulate
1 2 3 4 4
Ook & Omlűt & Nautli languages & Sperenjas
Papuan languages, Morphophonology, Lexical Semantics
Re: Alál: Vowels Exist
This is a great system. I really appreciate the time and effort you spent creating it.
- kiwikami
- roman
- Posts: 1203
- Joined: 26 May 2012 17:24
- Location: Oh, I don't know, I'm probably around here somewhere.
Re: Alál: Vowels Exist
Verbs and Queues and Argument Structure
or
I'm a Morphophonologist Don't Ask Me about Syntax
-----
Thanks, Creyeditor and spanick! Here's something that has nothing to do with vowels and is also a bit of a mess.
(I should be doing work right now so I thought I'd write this up rather quickly. It's unedited, unpolished, and probably slightly wrong, but it's a start.)
Alál's verbal agreement operates as follows. The arguments exist on a sort of queue, with two words that behave like prepositions performing a pair of operations on that queue (kxa appending to the end, and zta popping an argument off the front). In conjunction with the strict transitive-intransitive distinction and the use of the inverse voice, which reverses the two arguments that are actually marked on the verb, this produces a variety of conditions detailed below.
(1) Kıîtaa.
see<X.Y>-ACT
X sees Y.
This is a transitive verb, nothing special. It's in a form of the active voice that I refer to in my notes as "greater active" (contrasting with greatest, lesser, and least active) which is a distinction that isn't relevant here but mostly has to do with volition and some things regarding scope of the previously-discussed truth-value affixes. Basically, just note the final -a. The first argument in a transitive verb is necessarily the agent, while the second is the patient. The queue setup here as I have it in my notes is (X>Y), where the parentheses indicate which arguments are actually marked on the verb and the > indicates which argument is acting on which other one. Since there are only two arguments here, both are marked on the verb.
(2) Kıtàa.
see<X>-ACT
X is visible/X reflects light.
This is a basic intransitive verb with queue structure (X); again, nothing special here, except that Alal is fluid-s and in verbs with only one marked argument, the agent/patient status of that argument depends on the valency of the verb. Things typically don't get to determine whether they're visible or not, but for the sake of simplicity let's assume that X is capable of turning invisible at will, as all three of ACT, CAUS, and INV suffixes have four valency-based variations each and that's a bit much to deal with right now.
(3) Kıîtaık.
see<X.Y>-INV
X is (intentionally) seen by Y.
The inverse voice construction reverses the relationship of the two arguments of a transitive verb. It operates a little like a passive in that it promotes the object to subject position, but the valency is not reduced. This is a useful construction because the verb indicates the volition of the first argument, so it allows for nouns to which actions occur to have some input in the occurrence (as in “X got himself seen by Y”). The queue structure here is (X<Y).
(4) Kıîtaur.
see<X.Y>-CAUS
X causes Y to reflect light.
Here is the causative voice. The causative operates by adding a new argument slot, the causer, to the front of the queue. The verb will always interpret the two marked arguments as being in the first two available slots, so X becomes the causer and Y becomes what would be the subject of the intransitive verb. The queue structure is (X>Y). The causative voice can never appear unless valency is already at least two; that is, you can't get an intransitive causative construction.
(5) Kıîtaurık.
see<X.Y>-CAUS-INV
X is caused by Y to reflect light.
The causative and inverse can stack, flipping the two arguments; they must agree for subject volition. The queue structure is (X<Y). In general, assume the inverse can apply wherever there are two arguments marked on the verb; I won't provide examples for all cases. There is a single condition in which it can apply with only one argument marked, and I'll get to that shortly. It's fun.
(6a) Kxa Kıtàa {Z}
BEN see<X>-ACT Z.OBL
X produces light for Z's benefit.
(6b) Kxa Kıîtaa {Z}
BEN see<X.Y>-ACT Z.OBL
X sees Y for Z's benefit.
Enter kxa, the first quasi-prepositional operand thingie; because kxa acts as if it were a spatial preposition, its object gets displaced to after the VP, forming a discontinuous constituent (because I hecking love those). Kxa appends an argument to the end of the queue. In its most basic form, this argument will be a benefactive that operates like an indirect object or dative in a lot of constructions, and I generally gloss it as BEN. Other constructions (namely the causative) can force the kxa argument to take the patient role, but it will never do so outside of certain CAUS conditions; this is explained below. Kxa, unlike CAUS, can appear with only one marked argument on the verb, as we'll see later.
(7) Kıítkxaàha.
see<X.Y>-kxa-ACT
X sees Y, for someone's benefit.
Kxa may be incorporated into the verb, occupying the same position as the spatial axes (because again, it's essentially a preposition). In this case, the Z argument need not be overtly specified, though kxa still increases the valency by one. The queue structure is (X>Y) for {Z}, where the curly braces show that the Z isn't overt.
Note that kxa seems to act a little funny with verbs that aren't already marked for two arguments. That is, while it would make perfect syntactic sense for Kıtkxaàha (see<X>-kxa-ACT) to be equivalent to Kîtaa (see<X.Y>-ACT) but with an unspecified patient, meaning “X sees something”... that's not what occurs. Instead, what we get is the benefactive meaning, on an intransitive verb: “X reflects light for someone”. At first glance, this violates the rules set up for how kxa operates. It should add an argument in the next available slot of the role queue, which should be the patient. Kxa should mean the thing seen. But as it happens, the patient role... doesn't actually exist in this construction. It is added to the role queue by the presence of a transitive agreement marker. If the agreement marker is intransitive, as it must be for see<X>, then there is no patient role for kxa to fill. Thus, the next available role is the benefactive.
(8) Kxa Kıîtaur {Z}.
kxa see<X.Y>-CAUS Z.OBL
X causes Y to see Z.
Adding kxa to a causative verb works differently. Since the basic causative verb produces the meaning of X causing Y to perform some intransitive action, adding kxa...Z and increasing the valency by one makes that action itself transitive. In this case, the Z argument does take on the patient role, because the causative makes use of a two-part or “transitive” agreement marker (even though the two arguments marked are causer-agent, not actually agent-patient) and thus a patient role is added to the end of the role queue. In the normal causative, it's not filled or in any way pointed to, so the action Y performs is interpreted as intransitive, but add kxa and, even if the Z noun isn't specified, we suddenly have a patient. The queue structure here is (X>Y)>Z, where X and Y are marked on the verb and Z is not.
(9) Kıítkxaàur.
see<X.Y>-kxa-CAUS
X causes Y to see something.
Again, kxa may be incorporated into the verb, producing a (X>Y)>{Z} structure.
So how, one might ask, would we add a benefactive to this construction, since kxa is busy being very important elsewhere? How would we say “A causes B to see C for D's benefit”? Well, you wouldn't, not with only kxa. Here, we might use a perphrastic constructions, such as serial verbs “A causes B to see C – B sees C for D”. Or, if we don't mind the causer no longer being the subject... we can add in one more queue-shifting operation. But it'll be a bit before we get to this four-argument monstrosity of a construction.
(10) Zta Kıîtaa {W}.
zta see<X.Y>-ACT W.OBL
“X is seen by W for the benefit of Y”
First, enter zta. This preposition-esque-operand-thing inserts a new argument at the front of the queue. It takes the first role available to it, which here is of course the agent. This leaves X as the patient, and Y in the benefactive role, with queue structure W>(X for Y).
(11) Kıtztaàha.
see<X>-ARG-ACT
X is benefited by something reflecting light.
Using zta in this way is one method of promoting a benefactive object to subject position. This is straightforward in the intransitive; zta takes the agent role, and since it's an intransitive marker, no patient role is added to the queue, so the marker must refer to the benefactive. Here, we see it incorporated into the verb, just like kxa was able to be. As for promoting a benefactive such that it is the beneficiary of a transitive action, saying “X is benefited by something seeing something else”? Well, that's just the inverse of (10).
(12) Kıítztaàık.
ARG see<X.Y>-INV W
X is benefited by W seeing Y
Alright, moving on now to why I used to gloss zta as CAUSER (now just as ARG, because its usage is a bit more varied than that of kxa): It mostly shows up in causative constructions.
(13) Zta Kıîtaur {W}.
ARG see<X.Y>-CAUS W
X is made by W to see Y.
Here, the CAUS morpheme introduces a causer role that the first argument in the queue is going to take. That argument is the one introduced by zta, W. X and Y still fill the two verb slots, but as a regular agent-patient pair with the same relationship as in a non-causative transitive verb. The queue structure is now W>(X>Y). Since without the CAUS morpheme the role queue is only three slots long, a non-CAUS verb cannot have both kxa- and zta- marked arguments (which would bring the argument queue up to a length of four). But a CAUS-marked verb? That's just fine.
(14) Zta {W} kxa Kıîtur {Z}.
zta W kxa see<X.Y>-CAUS Z
W causes X to see Y for Z's benefit.
(15) Kıítztakxaàur.
see<X.Y>-zta-kxa-CAUS
X is made to see Y for something else's benefit.
And here we finally see that four-argument construction mentioned earlier. There's one more set of things to discuss now, and it involves the inverse. Recall that the inverse swaps the roles of the two arguments marked on the verbs. When there is only one argument marked, the inverse can still apply – as long as there is some oblique argument available to be swapped out. This argument must be marked on the verb; a benefactive given by the kxa...Z construction doesn't work.
(16) Kıtkxaàık.
see<X>-kxa-INV
X is benefited by something reflecting light.
An incorporated kxa works fine; it gives us the same promoted benefactive as in (11). Recall Kıtztaàha (see<X>-ARG-ACT) – the zta takes the agent role, and leaves X as the benefactive. Here, X would normally be the agent, and kxa indicates some benefactor, but INV swaps them, creating the same argument structure. There's really no reason to prefer one over the other, though I might come up with some semantic distinction later, save for the face that in this construction, kxa must be incorporated, while in Kıtztaàha, there's a zta Kıtàa Z unincorporated form where Z can be explicitly given.
(17) Kıtztaàık.
see<X>-zta-INV
X benefits something by reflecting light.
Here, zta fills the agent place, X fills the benefactive, and they swap, giving us the same argument structure as an active-voice intransitive construction with an incorporated kxa as in (6a). Again, no real difference except that in Kıtkxaàha, there's an unincorporated kxa Kıtàa Z version. The last way to get the inverse to work with intransitive agreement is also the most common way: promoting a locative noun to subject position. Recall that kxa and zta are prepositions, and their objects can be promoted when the preposition itself is incorporated into the verb in an axis position. So can other prepositions.
(17) Kıtùrûaık.
see<X>-inside-INV
Inside X, something reflects light.
Viola, a circumstantial voice of sorts. It's good fun.
(18) Kıtùrùztakxaàık.
see<X>-inside-ARG-BEN-INV
Inside X, something sees something else.
Here, kxa and zta break down a bit; they can be added here - both of them, which is normally disallowed outside CAUS constructions - and serve to specify the agent and patient of a transitive version of this. Kxa alone adds only a benefactive; both zta and kxa are required to force a transitive interpretation. Normally a patient role only appears when there's a transitive agreement marker, but out of sheer frequency of use for a transitive circumstantial compared to an intransitive circumstantial + benefactive, once the agent is specified with zta, if a kxa is already present, a wild patient appears. Structurally, (18) "should" be something like "Inside X, something reflects light for something else", but we simply don't get that interpretation.
Note that there's no way to get a causative circumstantial construction, or a transitive circumstantial that also has a benefactive. That's what serial verbs are for.
I'm procrastinating and shall stop now. I may come back to this later and clean it up / elaborate a bit, but I just wanted to write it down while I have the Alál notebook handy. My apologies for (6a) and (6b) breaking the numbering scheme; I realized I'd misnumbered them after already writing the rest of this. Might go back and fix that later.
or
I'm a Morphophonologist Don't Ask Me about Syntax
-----
Thanks, Creyeditor and spanick! Here's something that has nothing to do with vowels and is also a bit of a mess.
(I should be doing work right now so I thought I'd write this up rather quickly. It's unedited, unpolished, and probably slightly wrong, but it's a start.)
Alál's verbal agreement operates as follows. The arguments exist on a sort of queue, with two words that behave like prepositions performing a pair of operations on that queue (kxa appending to the end, and zta popping an argument off the front). In conjunction with the strict transitive-intransitive distinction and the use of the inverse voice, which reverses the two arguments that are actually marked on the verb, this produces a variety of conditions detailed below.
(1) Kıîtaa.
see<X.Y>-ACT
X sees Y.
This is a transitive verb, nothing special. It's in a form of the active voice that I refer to in my notes as "greater active" (contrasting with greatest, lesser, and least active) which is a distinction that isn't relevant here but mostly has to do with volition and some things regarding scope of the previously-discussed truth-value affixes. Basically, just note the final -a. The first argument in a transitive verb is necessarily the agent, while the second is the patient. The queue setup here as I have it in my notes is (X>Y), where the parentheses indicate which arguments are actually marked on the verb and the > indicates which argument is acting on which other one. Since there are only two arguments here, both are marked on the verb.
(2) Kıtàa.
see<X>-ACT
X is visible/X reflects light.
This is a basic intransitive verb with queue structure (X); again, nothing special here, except that Alal is fluid-s and in verbs with only one marked argument, the agent/patient status of that argument depends on the valency of the verb. Things typically don't get to determine whether they're visible or not, but for the sake of simplicity let's assume that X is capable of turning invisible at will, as all three of ACT, CAUS, and INV suffixes have four valency-based variations each and that's a bit much to deal with right now.
(3) Kıîtaık.
see<X.Y>-INV
X is (intentionally) seen by Y.
The inverse voice construction reverses the relationship of the two arguments of a transitive verb. It operates a little like a passive in that it promotes the object to subject position, but the valency is not reduced. This is a useful construction because the verb indicates the volition of the first argument, so it allows for nouns to which actions occur to have some input in the occurrence (as in “X got himself seen by Y”). The queue structure here is (X<Y).
(4) Kıîtaur.
see<X.Y>-CAUS
X causes Y to reflect light.
Here is the causative voice. The causative operates by adding a new argument slot, the causer, to the front of the queue. The verb will always interpret the two marked arguments as being in the first two available slots, so X becomes the causer and Y becomes what would be the subject of the intransitive verb. The queue structure is (X>Y). The causative voice can never appear unless valency is already at least two; that is, you can't get an intransitive causative construction.
(5) Kıîtaurık.
see<X.Y>-CAUS-INV
X is caused by Y to reflect light.
The causative and inverse can stack, flipping the two arguments; they must agree for subject volition. The queue structure is (X<Y). In general, assume the inverse can apply wherever there are two arguments marked on the verb; I won't provide examples for all cases. There is a single condition in which it can apply with only one argument marked, and I'll get to that shortly. It's fun.
(6a) Kxa Kıtàa {Z}
BEN see<X>-ACT Z.OBL
X produces light for Z's benefit.
(6b) Kxa Kıîtaa {Z}
BEN see<X.Y>-ACT Z.OBL
X sees Y for Z's benefit.
Enter kxa, the first quasi-prepositional operand thingie; because kxa acts as if it were a spatial preposition, its object gets displaced to after the VP, forming a discontinuous constituent (because I hecking love those). Kxa appends an argument to the end of the queue. In its most basic form, this argument will be a benefactive that operates like an indirect object or dative in a lot of constructions, and I generally gloss it as BEN. Other constructions (namely the causative) can force the kxa argument to take the patient role, but it will never do so outside of certain CAUS conditions; this is explained below. Kxa, unlike CAUS, can appear with only one marked argument on the verb, as we'll see later.
(7) Kıítkxaàha.
see<X.Y>-kxa-ACT
X sees Y, for someone's benefit.
Kxa may be incorporated into the verb, occupying the same position as the spatial axes (because again, it's essentially a preposition). In this case, the Z argument need not be overtly specified, though kxa still increases the valency by one. The queue structure is (X>Y) for {Z}, where the curly braces show that the Z isn't overt.
Note that kxa seems to act a little funny with verbs that aren't already marked for two arguments. That is, while it would make perfect syntactic sense for Kıtkxaàha (see<X>-kxa-ACT) to be equivalent to Kîtaa (see<X.Y>-ACT) but with an unspecified patient, meaning “X sees something”... that's not what occurs. Instead, what we get is the benefactive meaning, on an intransitive verb: “X reflects light for someone”. At first glance, this violates the rules set up for how kxa operates. It should add an argument in the next available slot of the role queue, which should be the patient. Kxa should mean the thing seen. But as it happens, the patient role... doesn't actually exist in this construction. It is added to the role queue by the presence of a transitive agreement marker. If the agreement marker is intransitive, as it must be for see<X>, then there is no patient role for kxa to fill. Thus, the next available role is the benefactive.
(8) Kxa Kıîtaur {Z}.
kxa see<X.Y>-CAUS Z.OBL
X causes Y to see Z.
Adding kxa to a causative verb works differently. Since the basic causative verb produces the meaning of X causing Y to perform some intransitive action, adding kxa...Z and increasing the valency by one makes that action itself transitive. In this case, the Z argument does take on the patient role, because the causative makes use of a two-part or “transitive” agreement marker (even though the two arguments marked are causer-agent, not actually agent-patient) and thus a patient role is added to the end of the role queue. In the normal causative, it's not filled or in any way pointed to, so the action Y performs is interpreted as intransitive, but add kxa and, even if the Z noun isn't specified, we suddenly have a patient. The queue structure here is (X>Y)>Z, where X and Y are marked on the verb and Z is not.
(9) Kıítkxaàur.
see<X.Y>-kxa-CAUS
X causes Y to see something.
Again, kxa may be incorporated into the verb, producing a (X>Y)>{Z} structure.
So how, one might ask, would we add a benefactive to this construction, since kxa is busy being very important elsewhere? How would we say “A causes B to see C for D's benefit”? Well, you wouldn't, not with only kxa. Here, we might use a perphrastic constructions, such as serial verbs “A causes B to see C – B sees C for D”. Or, if we don't mind the causer no longer being the subject... we can add in one more queue-shifting operation. But it'll be a bit before we get to this four-argument monstrosity of a construction.
(10) Zta Kıîtaa {W}.
zta see<X.Y>-ACT W.OBL
“X is seen by W for the benefit of Y”
First, enter zta. This preposition-esque-operand-thing inserts a new argument at the front of the queue. It takes the first role available to it, which here is of course the agent. This leaves X as the patient, and Y in the benefactive role, with queue structure W>(X for Y).
(11) Kıtztaàha.
see<X>-ARG-ACT
X is benefited by something reflecting light.
Using zta in this way is one method of promoting a benefactive object to subject position. This is straightforward in the intransitive; zta takes the agent role, and since it's an intransitive marker, no patient role is added to the queue, so the marker must refer to the benefactive. Here, we see it incorporated into the verb, just like kxa was able to be. As for promoting a benefactive such that it is the beneficiary of a transitive action, saying “X is benefited by something seeing something else”? Well, that's just the inverse of (10).
(12) Kıítztaàık.
ARG see<X.Y>-INV W
X is benefited by W seeing Y
Alright, moving on now to why I used to gloss zta as CAUSER (now just as ARG, because its usage is a bit more varied than that of kxa): It mostly shows up in causative constructions.
(13) Zta Kıîtaur {W}.
ARG see<X.Y>-CAUS W
X is made by W to see Y.
Here, the CAUS morpheme introduces a causer role that the first argument in the queue is going to take. That argument is the one introduced by zta, W. X and Y still fill the two verb slots, but as a regular agent-patient pair with the same relationship as in a non-causative transitive verb. The queue structure is now W>(X>Y). Since without the CAUS morpheme the role queue is only three slots long, a non-CAUS verb cannot have both kxa- and zta- marked arguments (which would bring the argument queue up to a length of four). But a CAUS-marked verb? That's just fine.
(14) Zta {W} kxa Kıîtur {Z}.
zta W kxa see<X.Y>-CAUS Z
W causes X to see Y for Z's benefit.
(15) Kıítztakxaàur.
see<X.Y>-zta-kxa-CAUS
X is made to see Y for something else's benefit.
And here we finally see that four-argument construction mentioned earlier. There's one more set of things to discuss now, and it involves the inverse. Recall that the inverse swaps the roles of the two arguments marked on the verbs. When there is only one argument marked, the inverse can still apply – as long as there is some oblique argument available to be swapped out. This argument must be marked on the verb; a benefactive given by the kxa...Z construction doesn't work.
(16) Kıtkxaàık.
see<X>-kxa-INV
X is benefited by something reflecting light.
An incorporated kxa works fine; it gives us the same promoted benefactive as in (11). Recall Kıtztaàha (see<X>-ARG-ACT) – the zta takes the agent role, and leaves X as the benefactive. Here, X would normally be the agent, and kxa indicates some benefactor, but INV swaps them, creating the same argument structure. There's really no reason to prefer one over the other, though I might come up with some semantic distinction later, save for the face that in this construction, kxa must be incorporated, while in Kıtztaàha, there's a zta Kıtàa Z unincorporated form where Z can be explicitly given.
(17) Kıtztaàık.
see<X>-zta-INV
X benefits something by reflecting light.
Here, zta fills the agent place, X fills the benefactive, and they swap, giving us the same argument structure as an active-voice intransitive construction with an incorporated kxa as in (6a). Again, no real difference except that in Kıtkxaàha, there's an unincorporated kxa Kıtàa Z version. The last way to get the inverse to work with intransitive agreement is also the most common way: promoting a locative noun to subject position. Recall that kxa and zta are prepositions, and their objects can be promoted when the preposition itself is incorporated into the verb in an axis position. So can other prepositions.
(17) Kıtùrûaık.
see<X>-inside-INV
Inside X, something reflects light.
(18) Kıtùrùztakxaàık.
see<X>-inside-ARG-BEN-INV
Inside X, something sees something else.
Here, kxa and zta break down a bit; they can be added here - both of them, which is normally disallowed outside CAUS constructions - and serve to specify the agent and patient of a transitive version of this. Kxa alone adds only a benefactive; both zta and kxa are required to force a transitive interpretation. Normally a patient role only appears when there's a transitive agreement marker, but out of sheer frequency of use for a transitive circumstantial compared to an intransitive circumstantial + benefactive, once the agent is specified with zta, if a kxa is already present, a wild patient appears. Structurally, (18) "should" be something like "Inside X, something reflects light for something else", but we simply don't get that interpretation.
Note that there's no way to get a causative circumstantial construction, or a transitive circumstantial that also has a benefactive. That's what serial verbs are for.
I'm procrastinating and shall stop now. I may come back to this later and clean it up / elaborate a bit, but I just wanted to write it down while I have the Alál notebook handy. My apologies for (6a) and (6b) breaking the numbering scheme; I realized I'd misnumbered them after already writing the rest of this. Might go back and fix that later.
Edit: Substituted a string instrument for a French interjection.
| | ASL | | |