Not All Irregularity Is Equal: Causally Isolating a Rare Failure Mode in Japanese Morphological Inflection

Excerpt

arXiv:2609.21179v1 Announce Type: new Abstract: Neural morphological generation systems often achieve high aggregate accuracy on benchmark datasets, yet such performance can conceal systematic errors clustered in rare morphological subclasses. We present an orthography-aware diagnosis of Japanese past-tense verb inflection, treating hiragana not merely as a transcriptional medium but as a representational system that encodes morphophonological structure. Using two character-level Transformer arc