It's finally here, logical sign language time! I've been posting about it on the LLL (and using my thread there as an idea dumping ground). More incomplete stuff is there, and for now I want to use this thread for slightly more complete stuff.
The project is an exploration into semantics and sign language based solutions to loglang problems. Some notable features include:
Basic form of word is the noun with verbs being reduplicands of nouns. Nouns can then be seen as 1-ary verbs (and be used for zero-copula constructions), and then verbs take various reduplication schema for aspect marking
SSM is done mostly prosodically with correction based on inherent syllable features
Slightly odd system of handshape marking
A novel zbalermorna based fingerspelling system used for foreign words
A class of zi'evla/freewords for sign names (this language will not have any native names) and country/culture names
SOV but still topic prominent word order (O^^/zo'u SV)
Potentially formalized systems of role shift and indexing
So, I'm having issues with the syntax and semantics, partly because of my lack of education but also I think because I'm just confusing myself 0_0
Essentially we're going off a neo-Davidsonian framework, with verbs taking just event arguments and nouns being specified for thematic role. Default forms of words are nouns, with extra movement signaling that the word is generated in the V head proper (I think). Modulation in that additional movement is treated as aspect, and as movement into the T head. Either way, the verb assigns theta roles to the rest of the sentence (probably ergative flavored).
That all feels well and good, but where I'm getting confused is on the vP and the correlation of thematic roles (semantics) to theta roles (syntax). I can't tell if I need a vP or not since tmk SLs generally avoid voice in favor of topicalization (movement to specCP). At the same time, I want some ergative-ish stuff going on and idk quite what implications that has on vPs with regards to unaccusative verbs versus unergative verbs.
Should I have a vP? Would it be useful for this ergativity, or can the ergativity just exist in the semantics of applicable verbs?
So if you wanna spell your name in Lojban, you can!
Considering making an adaptation to Toaq (maybe making one based on Derani?)
ETA: it would be very helpful for me to make this actually bc i can decide on more handshapes (cries)
Finally starting to get somewhere interesting! Deciding on what aspects I actually want to generate from movement 0_0
Basic movement yields the static aspect, basically a perfective or bare event.
Reduplication with a little loop makes a progressive verb.
Triplicating motion gives the durative, an "on and on" or repetitive sense.
More complex aspects include: Gradual continuous, which implies a slow speed of completion. Movement is slow and deliberate to match. Exhaustive, which is like za'o (going past natural endpoint). Signed like durative but maybe a different NMM.
An aspect I'm calling the screwedupitive, which is signed with shortened and tense motion. Unrealized, which is for incomplete actions, and signed with a shortened movement that doesn't quite finish. Completive, which might get a separate auxiliary.
And finally, my favorite one... Inconcinnative: the action is performed in an awkward and/or erratic way, signed with a stilted movement and some other associated NMM.
IDK how many of these I want to overlap or not. They could play decently nice together, so we'll see...
jan Olipija (of Luka Pona SL fame) gave me an interesting idea for 5uo5u: make it almost a taxonomy language
Basically, she thinks I should define general semantics for every handshape in the language, and then use those as initializations to derive new words. Entity classifiers would also fit in to this paradigm quite easily, as they can just stand in for the general category of that handshape.
My main questions are to do with unmarked handshapes, since those will be used a lot. I might not assign meaning to them at all, or allow for them to have many meanings. We'll see...
Handling classifiers will also be tricky, since idk how related I want them to be to their entity counterparts.
So yeah, what do y'all think of this kind of taxonomic vocabulary building?
The only problem I can see is that many words would look similar, leading to confusion, which is the same problem spoken taxonomic langs have. But I also know little to nothing about sign languages so feel free to ignore me.
No I think this could be an issue too, especially since if these work like sign initials then basically the vocabulary is like a matrix where part of the syllable means one thing and part means another, and that that creates the entire meaning.
It's like if you had a language with monosyllabic words where the onset was a general category and the rime was a different general category, and those combined to make the word.
Yeah I'm really not sure
Part of me thinks because it's an SL this will be less of an issue? But idk how true that is... SLs definitely do this kind of thing, whether with initials or with different "sound symbolism" type ideas (ASL kin terms, RSL internal anatomy all using the same shape, ASL's naked/bald shape, etc.), but it's not as absolute
Part of me also wonders if I could just do it less formally, like still have the general associations in each shape, but also have some irregularity and wiggle room.
Davidson's book on sign semantics is very useful. In some ways it feels like natural sign is more apt for being a loglang than speech. There's lots of little things she describes that give me ideas for various structures.
Scope of things like negation can be changed via what parts of a clause take a NMM:
[hs] PRENU JUNGAU-1 MI [/hs] {no prenu cu jungau mi}
PRENU [hs] JUNGAU-1 [/hs] MI {lo prenu cu na jungau mi}
Afterthought connectives:
ga lo broda gi lo brode > ^^BRODA rs(gi) ^^BRODE GA
Loci are equivalent to dynamic indices and can be used for specificity marking:
NANMU STASU NELCI+ {lo nanmu cu nelci lo stasu} "A man likes soup"
ix.a NANMU STASU NELCI+ {le nanmu cu nelci lo stasu} "A certain man likes soup"
ix.a ^^STASU NANMU NELCI+ {le stasu zo'u lo nanmu cu nelci} "A certain soup is liked by a man"
Adding some suffixes which specify the thematic role selected for nominalization. By default a noun is the event of the predicate, which for nouny unary predicates is just that noun (lo nu gerku = lo gerku here, basically), while for verby predicates it's like a gerund. But what if I want different nouns for each role of the predicate?
Inner case! Kind of like Nahaiwa or like the Lojban {lo SE broda} construction, but with a fun sign language twist. ASL's noun-verb pairs generally treat the noun as either an eventive or patientive without much care for the difference, with a special agentive suffix from the word PERSON. So, I decided to expand this system into a full inner case paradigm.
CITKA: A eats P
CITKA+ = {lo nu citka} "eating"
CITKA+pre = {lo citka} "eater"
CITKA+cpa = {lo se citka} "food"
I plan to make the PRENU and CPACU roots pretty simple as well so that compounding is very light and can keep syllable counts low. Maybe an open hand down the body for PRENU, and an open hand to fist at the chest for CPACU? That would make suffixing a pretty simple matter of adding an aperture node (maybe with additional path or setting changes too), since the selected fingers could be specified by the root word.
