A Constraint-Based Probe for Exploring Phonological Sound-Space Without Decipherment

This note describes an auxiliary, exploratory extension to a previously proposed phonological plausibility framework for evaluating unknown or ambiguous symbol strings. The extension introduces a phonological probe that generates broad sound-category constraints for structurally plausible symbol strings, while explicitly preventing phoneme assignment, lexical identification, or semantic interpretation. The probe operates by conditioning sound-category generation on phonological profiles, enforcing conservative phonotactic filters, and applying entropy-based failure detection to prevent overfitting. The method is falsification-oriented: its primary function is to rule out unsafe or unjustified sound hypotheses rather than to propose readings. Failure cases are treated as informative outcomes. This work is intended as a methodological demonstration for researchers in historical linguistics, undeciphered script studies, and comparative phonology, and makes no claims about any specific script or language. It extends the phonological plausibility framework described in *Basu, A. (2026). A Phonological Plausibility Framework for Evaluating Unknown or Ambiguous Symbol Strings*.