LUXVAR: A Constructed Language with Distinctive Morphemic Structure and AI-Recoverable Morpheme Families

Abstract

LUXVAR is a constructed symbolic language developed within the VARZIN research framework. This study presents a computational analysis of LUXVAR morphology, phonotactics, semantic structure tests, and AI-based clustering benchmarks. Main findings include: • Non-random morphological structure (z = 4.36) • Distinctive phonotactic fingerprint relative to multiple reference languages • Strong convergence of independent language models on morpheme families (inter-model ARI = 0.718) • No evidence for reliable recovery of the designed semantic axes from word forms alone The results indicate that the computationally recoverable structure of LUXVAR is morphemic rather than axial. This work is presented as a computational linguistics study. It does not claim empirical validation of cognitive, psychological, or physical effects associated with the language. The remaining open question is whether human observers recover the same morpheme-family organization identified by independent language models, a question reserved for the proposed VPE-001A human study.