Toward Automatic Discovery of a Canine Phonetic Alphabet

Dogs communicate intelligently through vo-calizations, yet the phonetic properties of their communication remain poorly understood. This paper introduces an iterative algorithm inspired by human phonetic discovery, leveraging minimal pairs to identify distinct canine cognitive vocal units and construct a comprehensive phonetic alphabet. Additionally, the algorithm derives canine vocal pattern that exhibit structured correlations with specific environments and activities, suggesting potential meaningful communicative patterns. Our approach provides a novel framework for analyzing non-human vocalization systems and offers potential applications beyond canines to other animal species, advancing the study of cross-species communication.