Creyeditor wrote: ↑10 Dec 2023 09:07
Salmoneus wrote: ↑08 Dec 2023 16:06
My conlang, U, departs from normal linguistic (Saussurean) expectations by not using signs that represent things or ideas. [at least, not at the heart of the language; there are some functional gestures that could be seen, I guess, as symbolic]. I never got too far into it, though (partly because I got bored making diagrams).
How does that work? Is it the pairing non-arbitrary?
There is no pairing! Or, to put it another way, the linguistic gesture is not one of the things that is paired.
So, there's actually two different ideas of 'U', which work differently, but with a core similarity: the 'original' idea I had, and wrote up a bit about a few years ago, and an alternative idea, which is maybe more radical but which I haven't explored in any detail at all. I mean, I haven't explored either in any detail, but I haven't explored the second at any length, even, so it might not work, or might not feel 'right'.
---------
So, the original idea...
U does have 'words', but they're not directly expressed in language. They're more like 'concepts'. In Human, we can think of 'concepts' being arbitrarily and (semi-)uniquely, absolutely, 'labelled' with (mostly) fixed linguistic gestures: 'horse' is the label for the concept of the horse, which refers to actual horses. [whether this is philosophically or psychologically accurate as a model of how Human works is debateable, but it's a useful story to use for contrasting with U].
But in U, concepts are not labelled absolutely, but relatively. We can imagine a speaker of Human having a huge, unordered pool of concepts in their head, and they have to pluck out the right one by grabbing hold of a tag attached to it.
In U, there are no tags. But, to compensate,
the pool is not unordered. Instead of a pool, it's a complex net of 'positions', linked to one another in a certain pattern. Each concept occupies a place on the net. [the net is probably fractal and multidimensional]
So, to 'access' a concept, the speakers 'physically manipulate' an imagined net with their arms, to move from the place where one concept is stored to a place where another concept is stored. Then they pluck out that concept, that word, and place it in a 'sandbox', as it were, of concepts. The concepts remain in that sandbox throughout the conversation, and can easily be re-accessed by simply 'physically' picking them up and using them. And differen't speakers, if close enough, can manipulate words in each other's sandboxes (or on their 'table', perhaps would be a simpler metaphor, if less aquatic). There are then functional gestures to indicate how one word is being used with the other, which I believe takes the form of effectively 'attaching' words to one another.
So, for example, to say "shark ate clownfish", you use your (many, many) arms to navigate through the imaginary net from an 'access point' (chosen/defined through the pragmatic context of the conversation, loaded with implications about politeness and purpose) to the place where 'shark' is stored, and also the places where 'eat' and 'clownfish' are stored. Possibly simultaneously, depending on how many arms you have free. You take those concepts and put them on your table, and then you enact attachin 'clownfish' and 'shark' to 'eat' in a way that distinguishes subject and object through location/orientation. This creates a compound concept of a shark eating a clownfish, and then there's probably a further way of moving that concept, or gesturing to it, that signifies the past tense.
Excitingly, speakers can in reverse place concepts from the table - taken from elsewhere in the net, or from the interlocutor's net, or summoned deictically from the environment - back into the net at a new location. This can create new vocabulary, and results in the same (or very similar! with scope for differences in nuance!) concepts being found in multiple locations, which can slash navigation time from concept to concept.
[we might at first assume the net was arranged logically, with similar concepts near to one another. It's not. Items are arranged so as to minimise navigation time/space between things that are likely to be involved in the same expressions - so 'shark' and 'eat' are probably very close!]
But the exciting thing about that is:
there's only one U language for all speakers across the globe! But it's not mutually intelligible. What happens in effect is that speakers of nearby 'dialects' chiefly remain within certain adjacent parts of the net, while speakers far away may occupy very different parts - parts so remote that the different speakers may be unfamiliar with them. But as you move from community to community, you move linguistically as well as geographically, but adjacent dialects remain intelligible because they simply use adjacent, overlapping areas of the net. [Likewise the net evolves over time]
[this is probably an exaggeration; there may also be different 'accents' in how, for instance, you symbolise your navigation of the net. But the differences I think would be very small compared to human linear syntax variation]
Oh yes, if it's not clear, U isn't exactly linear. Obviously conversations take time so there's an inherent linear dimension. But because concepts are also arranged on the table in (imagined) space, they can be 'used' non-linearly or even simultaneously - you can talk about a shark eating a clownfish over HERE, while simultaneously (thanks to your many arms!) talking about geopolitics over at THIS part of the table. You can even 'steal' words and sentences that have been left lying on an adjacent conversation! And while you're using concepts on the table, you can still be finding new ones in the net. Unlike in Human, "saying" a word and "using" it aren't the same action - you can say a word long before you actually use it. [the pragmatic nuances of the words left lying around unused during a conversation must be immense! Where a human might talk positively about a popular politician and then go "COUGH dictator COUGH" to indicate their real feelings, U speakers can simply discreetly place 'dictator' in a corner of their table the entire time... and they probably do this with multiple things ALL tHE TIME. (eg there will be constant choices about which words and sentences to leave 'lying around' as the conversation moves on, and which to discard)]
Obviously, speaking U requires phenomenal powers of spatial visualisation and memory. This is a feature - I wanted somethin humans couldn't speak even if they wanted to. [and U speakers might say that a human rote-memorising maybe 40,000 near-random, unconnected 'words', and using them in speech in an entirely linear way that required speakers to mentally access things said long before but without any spatial cues or even temporal reference marks to act as reminders, was an unrealistic thing...]
-------------
This, however, may be a bit too loglangy and human.
So more recently I've thought of a modification: what if words aren't one position on the net, but a PAIR of positions? This greatly increases the storage density of the net. It also inherently embeds a concept of argument structure - subject and object could be 'placed' onto different 'ends' of the word (some words may even include more than two points). And it further breaks the sign/signifier pairing, since there would be navigational choices even WITHIN the articulation of a word (moving from one end to the other). The 'use' phase could then be, in effect, the construction of a parallel, temporary net. (or more than one!) But I've not really thought through the implications of this idea.
[you could go further and make the word the specific ROUTE between two or more points, which would further increase density of storage. But it would restore an element of the sign/signifier pairing, and feels less exciting]