Mayangoli errors are context-sensitive errors in Tamil that arise from confusion among phonetically similar graphemes (e.g., ல/ள/ழ, ர/ற, ந/ன/ண). These errors are challenging for conventional spell checkers because both incorrect and correct forms are valid dictionary words, making dictionary lookup insufficient and requiring contextual modelling. We present TamilMayangoliSpell, a reproducible framework for Mayangoli error correction that combines (i) Tamil-specific preprocessing for sentence segmentation and normalisation, (ii) linguistically grounded error induction for generating training data constrained by dictionary validity, and (iii) fine-tuning of multilingual sequence-to-sequence models. Using 30,000 sentence pairs derived from TamilCorp, a massive multi-genre Tamil corpus and split 80/10/10 into train/validation/test, we fine-tune mBART, mT5, and NLLB under a small hyperparameter grid using greedy decoding with a maximum sequence length of 128. mT5 achieves the best performance (BLEU 99.28; Exact Match Accuracy 93.50%) and remains strong in a cross-genre evaluation on short stories. The preprocessing scripts, generated parallel datasets, and trained models are publicly available in a GitHub repository.

As part of a project to construct an interactive program which will encourage children to play with language by building jokes, we have developed a large lexical database, closely based on WordNet. As well as the standard WordNet information about part of speech, synonymy, hyponymy, etc, we have added phonetic representations and symbolic links allowing attachment of pictures. All information is represented in a relational database, allowing powerful searches using SQL via a Java API. The lexicon has a facility to label subsets of the lexicon with symbolic names, and we are working to incorporate some educationally relevant word lists as sublexicons. This should also allow us to improve the familiarity ratings which the lexicon assigns to words.