# Mind Your Moras: Orthography-Aware Error Analysis of Neural Japanese Morphological Generation

arXiv: https://arxiv.org/abs/2605.20043 · HTML: https://mlsanity.com/paper/2605.20043/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 15 references that are in the map.
- First posted on arXiv 19 May 2026.

## Builds on (most cited first, 6 of 6)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Synthetic and Natural Noise Both Break Neural Machine Translation](https://mlsanity.com/paper/1711.02173/) (2017)
- [Information-Theoretic Probing for Linguistic Structure](https://mlsanity.com/paper/2004.03061/) (2020)
- [Reporting Score Distributions Makes a Difference: Performance Study of LSTM-networks for Sequence Tagging](https://mlsanity.com/paper/1707.09861/) (2017)
- [Exact Hard Monotonic Attention for Character-Level Transduction](https://mlsanity.com/paper/1905.06319/) (2019)
- [Imitation Learning for Neural Morphological String Transduction](https://mlsanity.com/paper/1808.10701/) (2018)

## Built on by (most cited first, 1 of 1)

- [Not All Irregularity Is Equal: Causally Isolating a Rare Failure Mode in Japanese Morphological Inflection](https://mlsanity.com/paper/2609.21179/) (2026)

## Abstract (arXiv)

We present an orthography-aware error analysis of Japanese past-tense morphological inflection, treating hiragana not merely as a transcriptional medium, but as a representational system encoding morphophonological distinctions that may influence model generalization. We evaluate two character-level sequence-to-sequence architectures on past-tense formation using datasets formatted according to the SIGMORPHON 2020 and 2023 shared task conventions. Despite high aggregate accuracy, models exhibit systematic, linguistically interpretable errors that cluster around specific orthographic properties of hiragana. We introduce a concise error taxonomy capturing seven primary failure modes and provide both quantitative and qualitative analyses. Gemination-related errors dominate residual failures, accounting for 75-80% of errors, particularly in verbs whose stems end in the vowel e and require gemination before the past-tense suffix. Error patterns remain highly consistent across architectures and random seeds, suggesting a robust interaction between orthographic representation, morphological structure, and data frequency effects in shaping model generalization. These results underscore the necessity of orthography-aware evaluation for understanding neural generalization in morphologically complex languages.