- εἰμί - I am
- εἶ - you are
- ἐστίν - he/she/it is.
The breakthrough came when I heard William A. on Conlangery mention that many languages families (Athabaskan and Uto-Aztecan for example) encode both the subject AND THE OBJECT in the verb! So, for example, one word in Navajo:
- yishạ́ - I am eating it.
This gained me a lot of milage, but it felt a bit like cheating. Aren't I just hiding the (pro)noun, not eliminating it? How could I get rid of it more completely? As I thought about it, I remembered that 'here' and 'there' are adverbs, so if I could replace person in the encoded subject/object with demonstratives, I would be closer to my goal.
Deixis is the over-arching concept behind 'this' vs. 'that'. My experience with Korean taught me that there are (at least) three levels of deixis:
- This. By me. Proximal. Here. 이사람은/this person
- That. By you. Medial. There. 그사람은/that person
- Yon. Not by either of us. Distal. Yonder. 저사람은/yon person
As you can see, English previously had three levels, but has generally lost this 3-way distinction in favor of a two-way one. Sad. R.I.P. 'yon'.
So far, Perelandran would then be able to say, in one word, 'that-hit-this', which could mean 'you hit me' or a host of other things.
Next, I discovered the Seri language, which has nine demonstratives that are mixed up with definiteness ('a' vs. 'the'). I've decided to go with eight in Perelandran:
- Definite, Proximal, standing Up (DPU) - "this standing person"
- Definite, Proximal, Sitting (DPS) - "this sitting dog"
- Definite, Proximal, lying Down (DPD) - "this mat which is lying here"
- Definite, Medial, Coming (DMC) - "that approaching storm"
- Definite, Medial, Going (DMG) - "that woman who is walking away"
- Definite, Distal, Coming (DMC) - "yon attacking (us) army"
- Definite, Distal, Going (DMG) - "the horse riding off into the sunset"
- INDefinite (IND) - "some people"
- Persons/Lightning/(dynamic) Forces (PLF) - uncontrollable, (mostly) unpredictable, intentional things. cp. "acts of God/Nature"
- Living/Animate Beings (LAB) - more predicable, but very alive. Cows, cats, and gravity.
- Receptacles/Openings/Controls (ROC) - that which is manipulated. An element of unpredictability remains, since anything can go in a box.
- Inanimate objects -- predicable and controllable -- are all at this level, but subdivided into three sub-categories:
- Solid/Discreet Things (SDT) - balls, fruits, non-containers with clear boundaries
- Mushy, Porous Areas (MPA) - fuzzy boundaries, fuzzy definitions, e.g. mud or a mirage/reflection on a hot day.
- Bendy, Flat Fractals (BFF) - On Perelandra, "land" is actually great clumps of tough "seaweed" growing together into "islands". This categories sees the interlocking, patterned nature of bushes, carpets, rope and shower curtains together.
- Flowing/Undulating Flocks (FUF) - a school of fish is not the sum of individuals. You can't study individual water molecules and understand a river.
- Cloudy, Random Hazes (CRH) - There is continual, complete cloud cover on Perelandra. It's always moving and changing, but means nothing.
- Unchanging, Dead Stuff (UDS) - This is how we see a majority of things in our world, whereas Perelandrans consider this the least populated, least important category there is.
On our world, we 'name' everything. On Perelandra, only PLF's get names. If you talk to your dog, you've promoted them to a PLF. But even people progress down through the levels. Babies are LAB's. When a person is a sleep, they become ROC's, contains for dreams (which are very important on Perelandra). A corpse is an SDT, but quickly decomposes into a MPA, then a BFF (!) and finally, because everything is buried "at sea", a FUF. Once the deceased memory has begun to fade in the collective subconscious, they are a CRH. The 'are not now' are distinguished from the 'never have been' as CRH's vs. UDS's. If someone has no idea what a pencil is for, it's a SDT. Once they learn, it's an ROC.
Hi Robert. I enjoyed your post, and I wanted to make some comments. I don't know that your scheme for replacing pronouns with deictics really does what you think it does; i.e., gets rid of pronouns. It seems to me that you're just replacing one set of markers with another set that happens to be homophonous with the set of deictics, whose interpretation is identical to pronouns. Something like this has happened in many languages, though not to this extent. Typically you find deictic elements replacing 3rd person pronouns: this happened in Latin and Shoshone. Replacing 1st and 2nd persons with deictics is more unusual.
ReplyDeleteIn a project of mine, I reconceptualized pronominal reference as vectors. This is clearest with transitive predicates. I have a set of proclitics which "point" the action in a particular direction, with the directions being grammatical person. This idea is clearest with transitive verbs. Here are the clitics:
le= second person acting on a first person (2 > 1)
ku= second person acting on a third person (2 > 3)
wa= first person acting on a third person (1 > 3)
0= third person acting on a different third person (3 > 3')
So from the stem kipe 'poke', you get the following short paradigm:
lekipe 'you poked me'
wankipe 'I poked him'
kunkipe 'you poked him'
ikipe 'he poked him'
(The n- is a transitive marker; it alternates with i- in word-initial position.)
The direction of these "vectors" can be inverted with the inverse marker l-, which takes the place of the transitive marker. (l- alternates with u- word-initially.)
lelkipe 'I poked you'
walkipe 'he poked me'
kulkipe 'he poked you'
ukipe 'he got poked by him' (not really a passive, but it's the best English can do)
So that's a different way to do it, and I think it genuinely gets rid of person agreement. Person information is still available, but as part of a "vector" and not as a discrete referent.
Thank you so much for coming by, Taipo. I really appreciate your comments: it's so nice to hear from a professional!
ReplyDeleteI had heard about the "vector" things from ... was it Navajo? Some language have this hierarchy, where 2>1>3 '>3 or something like that. If you go "against the grain" you have to mark the verb. I've pretty much decided that there will be no objects objects, that to have a D.O. or I.O. you'll need to restate the verb in another voice.
I think you're right, though, that if all I did was replace pronouns with deictics, that's not actually eliminating much. If on the other hand, things could be conceived of as "source" or "vector", then that would be closest to pure "verbiness". Ex.:
blow-0.VALENCY, destroy-1.VALENCY-PAST-PERFECT-DPU-BFF dwell.in-1.VALENCY-NON-FINITE
This would mean literally, "There was wind and this matted-standing-dwelling has been destroyed" but more naturally in English, "The wind destroyed my house."
What do you think?
Algonkian has the person hierarchy 2 > 1 > 3 > 3' and inverse constructions, which is where I got it from. So my l- prefix is nothing more fancy than the Algonkian inverse. However, I don't think that any natlangs conceptualize person marking as a vector; Navajo and Algonkian both mark subjects and objects separately as discrete arguments of the predicate (though Algonkianists would probably want to argue the point with me; it's a little more complicated than that).
ReplyDeleteI don't think we can get away from dividing the world into entities, events, and qualities conceptually, but I think it is possible to erase the distinction *grammatically* by inflecting all stems alike. That's what my project[1] does: in this way, I have been able to get rid of nouns, if only because all entities are expressed in predicates meaning "be an ENTITY". If entities are expressed as stative predicates, they are then eligible for tense/aspect marking, person vectors, and so forth. Your example looks like it's moving in that direction.
One other feature that comes in handy in my project is the use of what I call determiners. Determiners introduce subordinate clauses as adjuncts to the main clause. Grammatically speaking, the only obligatory clausal element is the predicate; everything else is optional. To include other elements, they need to be introduced by a determiner. So your sentence would be conceptually translated as:
3>3'= destroyed DET 3=be_wind DET' 3=be_house
It destroyed it'; that which is the wind, that' which is the house.
The determiners track which main clause argument they are coindexed with, which I'm showing here with the ' symbol. It's sort of like lambda abstraction from formal logic (or what I remember of lambda extraction from grad school).
[1] Miapimoquitch, formerly Tepa; Tepa was profiled in Conlangery #8, but I have yet to put Miapimoquitch online.
I blogged about your awesome questions:
ReplyDeletehttp://reformedconlanger.blogspot.com/2012/06/objectlessnessish.html