Software
or, PhonoLeg-BR. Confidence-based phonotactic scoring derived from TINA's internal neural decoder states.
Enter a word above and press Score.
The phoneme sequence is familiar to the phonological grammar of Brazilian Portuguese. The model decodes with high confidence at every position. Suitable candidate for psycholinguistic stimuli or neologisms.
Marginally licit sequence. May contain atypical consonant clusters, infrequent vowel patterns, or unusual syllable structures. Useful for studying the gradient zone of BP phonotactics.
The sequence violates core phonotactic patterns of BP. High internal entropy in the decoder indicates that no familiar phonological mapping was activated. Likely perceived as a non-word by native speakers.
PhonoLeg-BR estimates how phonotactically natural a word or pseudoword sounds in Brazilian Portuguese. It does this by measuring how confidently and consistently TINA — the underlying neural model — processes the input: words with native-like sound patterns are handled smoothly and receive high scores, while unusual or foreign sequences cause the model to struggle, yielding lower scores.
The IPA transcription shown alongside the score is a by-product of the analysis, not its basis. The score reflects the stability of the model's internal process across multiple passes — the more consistent and confident that process, the more phonotactically native-like the word is considered to be.
Underlying model: TINA (Barroso, 2026) — hybrid neural-symbolic system trained on 95,927 orthography–transcription pairs of Brazilian Portuguese.
This tool is still under active development and must be used cautiously.