Whale Song is a Language Precursor with Context Free Grammar

Previously, no evidence existed for the early stages of language evolution. Here, the recursive ordering rules of hump back whale song—encoded in the mutually exclusive periodic patterns of the fundamental frequency sequence—is decoded, revealing that it constitutes a highly ordered pattern of symbol system, far beyond what has been previously discussed. A hierarchical unit of ‘word', an intermediate division between the song and theme hierarchies, has been identified. It is shown that the symbol sequence has automorphic syntactic structures between hierarchical levels, and that the complexity of song is the complexity of words. Phrase boundaries are encoded using formal symbol markers and the principle of prosodic descent. All three of the following conditions are satisfied: 1) the recursive application of automorphism rules, 2) the presentation of phrase structure through the alternation of content words and formal symbols, 3) the presentation of phrase boundaries through F0 re-rise-fall patterns. These new findings consistently conform to the formal requirements for grammar to emerge in human language and be learnable. That is, evidence is presented that humpback whale song is a type of language precursor belonging to the computational class of context-free grammars. Furthermore, the pitch distribution of sound units, the minimal structural unit of song, is concentrated around the consonant intervals perceived by human hearing with simple integer ratios—octaves and perfect fifths. These results suggest that language and tonal music share a common origin, and that symbolic systems are less a product of transmitter-specific constraints, and more as a consequence of the universal structural evolution of the information itself.