Kingfisher - An Experiment in Language Evolution

Kingfisher is a a small experiment to see how languages evolve, given a particular biome and communities.

Seeding:

N communities are distributed between various biomes across the generated continent(s)?

Each community is seeded with 200 words (100 nouns, 50 verbs, 50 other). Using a genetic algorithm, each community evolves it’s language down to 100 words, which then results

Notes:

[#A] Mapping generation

Polygonal Map Generation for Games - http://www-cs-students.stanford.edu/~amitp/game-programming/polygon-map-generation/

Starts with graph structure and builds outwards

Desert Generation - https://github.com/mewo2/deserts

Generating fantasy maps - http://mewo2.com/notes/terrain/

Random map generation - https://gamedev.stackexchange.com/questions/31241/random-map-generation

Dwarf Fortress map generation: http://www.gamasutra.com/view/feature/3549/interview_the_making_of_dwarf_.php?page=7

Geometric tesselations?

n-gonal generation?

Voronoi Tesslations? - http://www.gameprogrammer.com/fractal.html

Coastal generation/contientents/etc

Civ/endless legend/4x mapping generation?

Glacial processes: http://uregina.ca/~sauchyn/geog323/glacial1.html

Notes

Uses buildings, but a simple guide: https://www.cs.auckland.ac.nz/~roboteam/documents/RandomMapGenerator.pdf

Lynda: https://www.lynda.com/Unity-tutorials/Unity-5-2D-Random-Map-Generation/475939-2.html

[#B] Language development

Language Evolution

Language Evolution Simulation - https://fatiherikli.github.io/language-evolution-simulation/

Uses agent based evolution, we’ll use cross pollination instead?

https://www.wikiwand.com/en/Agent-based_model

Neural networking to evolve languages

Languages as base?

Arabic - Koran/?

Chinese - Journey to the west/etc

Latin - Romance languages - plato’s works/homer’s works

Migration: https://en.wikipedia.org/wiki/Human_migration

Interactions with each other - neural networking - evolves the corpus with each other’s works - markov chains???

Markov Chains Based Random Word Generation: http://ben.akrin.com/?p=779

Another: https://terse-words.blogspot.com/2012/02/generating-gibberish-in-python-markov.html

Skip-gram model with fasttext/gensim

AI Lectures: http://ai.berkeley.edu/lecture_videos.html

Model generation?

Natural Language Inference Chains: https://arxiv.org/abs/1606.01404

Generating naming languages: http://mewo2.com/notes/naming-language/

Note that: ‘The initial idea was that different regions of each map would have different languages, but I abandoned this because it was too hard to make it clear that this was what was going on…’

We solve this problem by having different character bases

HN discussion: https://news.ycombinator.com/item?id=12186022

Gutenburg based gibberish in python: https://github.com/colwilson/Gibberish

Notes

Extra papers: http://www.langev.com/author/ksmith

NLP Professor: http://www.linguistics.illinois.edu/people/girju

Language and Social Interaction: http://www.linguistics.illinois.edu/people/koshik

Lingusitics Professors: http://www.linguistics.illinois.edu/people/faculty/default.aspx

Evolution of Human Languages Literature: http://groups.lis.illinois.edu/amag/langev/pubtype/inbook_TheEvolutionofHumanLanguages.html

Arbitray Communciative Conventions: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0043807