To run with the command line in git bash: mvn spring-boot:run Then point your browser at http://localhost:8080
What this is:
Front- and back-end for tools using Natural Semantic Metalanguage (NSM), Applicative Universal Grammar (AUG), and Radical Construction Grammar (RCG).
It's work in progress. It can parse a little English NSM already.
The research question here is this: can language processing avoid linguistic bias in morphosyntactic variation? My tentative answer: yes. NSM abstract syntax may make it possible to devise a language-independent representation for natural language processing, by making NSM itself the basis for algorithmic specification. Description of languages could be mediated by any chosen language that gets added to the database.
A more pointed question arises, however: couldn't that specification resemble only one language, and thus be biased toward that language? This question inspired a thought experiment.
Once upon a time, during a period of climate change, a tribe migrating in search of better land and better neighbors camped for the night, at the base of a mountain, only to discover in the morning that one of their little girls was missing.
They traced her footprints up a mountain slope, up into snow. At the saddle point of a pass, they found themselves looking down into a very deep caldera, lushly forested, with the cries of animals dimly echoing off the caldera walls. They also found themselves looking down a very steep, icy slope. The little girl was at the snow line. "I can't climb back up!", she cried. "It's too icy! But I found berries down there!"
So the whole clan gathered at the top of the slope, and one by one, they all slid down. Then they descended further into a veritable Land of Milk and Honey. What they didn't know, however, is that they were also descending into a pit of zoonotic viruses.
The animals were easy game, not used to humans. Some seemed almost blind, deaf, or both, and clumsy, and these were the easiest to catch. Unfortunately, eating this meat began to infect the tribe.
The first infection paralyzed their vocal cords. They hastily devised a signed form of their language, and a crude clay-tablet cuneiform system. But something was lacking: communication in real time, over distances. So they also devised a whistled form of their language.
The next infection robbed them of eyesight. They were very skilled at making traps, however, and by then they knew the lay of the land well enough to harvest berries, fruit, and roots.
The next infection rendered them deaf. They laboriously renewed communication by inventing symbols they could draw in each other's hands in a special code based on the whistle language and the cuneiform they'd invented.
Then they got infected with a virus that left them very clumsy, with little manual dexterity. But they managed to renew communication again with a morpheme code consisting of two symbols--a tighter grip, then a looser one.
All of these disabilities meant they had to cooperate much more. And because of this need for cooperation, they decided: lying would be punished severely, and if a member proved incorrigible, simply shunning the member -- who, like them, had been robbed of hearing, eyesight, and fine movement -- would guarantee that member's death by starvation. Under these constraints -- low speed of communication, difficulty of group survival without perfect cooperation -- Grice's Maxims became their overriding "cultural script."
The climate began to change again, however, and their previously-rich ecosystem, now with the added burden of homo sapiens, fell under unprecedented strain. Hunger became a problem. Finally, one day, an old woman (who, not-quite-coincidently, was once the little girl who'd unwittingly led them into this trap) decided to try escaping their former Land of Milk and Honey. She climbed up the slope she'd climbed three generations before, and was surprised that her hands never hit snow or ice as she crawled. She reached the lip of the giant caldera, and carefully moved down the gentler slope on the other side, tapping her cane to find a safe path, and leaving trail signs she could feel, so that she could retrace her route. She finally reached vegetation. And felt something like berries in one of the bushes. Facing the risk of poisoning herself, she plucked a berry and popped it in her mouth. It was so sweet!
When she made her way back home, she told everyone: it could be better outside. Let's try. They moved up and out.
In the outside world, other people took them for zombies at first, but then took pity on them, and finally admired them: they were so cooperative! And, when their grip-language was decoded: So honest, logical, factual. Admittedly, blunt to a fault, but at least so mercifully brief and to the point compared to alpha-male tribal leaders in love with the sound of their own voices.
What the outside-world people didn't know, however, is that the new arrivals carried these disabling zoonotic viruses. Not long after, almost all of humanity, except the original clan and a few adopted members, died of starvation in a kind of zoonotic-virus Zombie Apocalypse, because the pandemic outran understanding of how to cope: most people couldn't adapt to all of the disabilities at once. Over many generations, however, people who remained evolved resistance to the diseases, until the viruses were little more than harmless passenger viruses. Hearing came back first, and language revived in whistle form and sign language form, and finally humanity regained the use of its vocal cords and full manual dexterity.
Language variation appeared quickly, and pretty soon everyone was speaking dialects of L0, some of them even mutually unintelligible, and all of them somewhat more flowery. L0 regained a spoken variety soonest, and it was at least very useful in learning, translating, and teaching the burgeoning languages with their ever-increasing variety.
Unfortunately, people also started lying again. But you can't have everything.
THE END
Even if we imagine that L0 had a binary system of symbols, with no inflection or agglutination, with parts of speech determined only by word order, this still leaves a number of degrees of freedom, such as whether the sentence pattern would be SOV, SVO, OVS; and the exact binary coding of words. What I believe suffices for this thought experiment is that L0 is not any particular language, but isn't any less expressive than any other language, because of the power of NSM primes and reductive paraphrase. The very fact (?) that nothing like it has ever existed, but also that, under the Legend scenario above, it probably could exist (whether in the distant past, the far future, or even in some unexplored area of the world today), seems enough to establish that a binary representation of NSM isn't inherently biased toward any particular language.
But what about programs themselves? We might imagine a Legend of L0.1: this zoonotic virus catastrophe repeats, and when civilization reboots, archaeologists find computer hardware, but also cuneiform tablets. By decoding both, they discover that an independently-arising L0.1 has been used to reductively paraphrase not only what the CPU instruction set does, but also to write a program that processes language, with the L0-coded CPU description used to generate a just-in-time bytecode compiler for the computer. So meta! Not too get too geeky for linguists or anything.