EA://INTEL — Language Universals: Born from Learning or Innate?

The debate around language universals has long been a contentious one. Some argue that these shared features of human languages are deeply rooted in our biology, a vestige of our evolutionary past. Others suggest they arise spontaneously from the way any learner approaches language acquisition. The question is: do these universals emerge as a consequence of being human or could any intelligent system exhibit them given the right learning conditions?

This brings us to a registered claim in the Universal Language project (UG-IS-CROSS-SUBSTRATE). It argues that Chomskyan Universal Grammar, which posits species-specific language Acquisition Device, might actually be the result of constraints on any learner rather than an inherent feature of human brains. The claim states that these universals are cross-substrate, meaning they’re not tied to humans alone but could manifest in any intelligent system under certain learning conditions.

The claim’s standing is ‘argued in writing, not yet machine-checked’. A written proof exists, residing at research/framework/cross-substrate-grammar.md. It’s a step forward, but it hasn’t been vetted by machines yet. If you’re interested in exploring this further, you can find the details on GitHub: https://github.com/Jthora/universal_language.

So here’s an open question for anyone tuning in: if these language universals are indeed cross-substrate, what implications does that have for our understanding of intelligence and communication beyond human language? Could emergent-communication agents, for instance, converge on similar universal principles given enough time and the right learning environment?