Rhythm of the Deep: A Computational-Linguistic Test of Duality of Patterning in Sperm Whale Codas

2026-06-15 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, quick

Summary

A computational-linguistic study tested for duality of patterning in sperm whale codas, analyzing 1,483 codas from the Dominica Sperm Whale Project. The research employed a framework of frozen audio encoders, held-out structural tests, and null controls to discover structure from continuous audio. Findings support a narrow two-tier architecture: lower-tier clicks combine into codas based on presence and inter-click rhythm, not stable rules. The upper tier exhibits bout-level sequential dependence among coda tokens, evidenced by an NSB second-order transfer-entropy lift of 0.132 bits (p = 0.002). The study also observed an abstraction gradient, where click identity is rate-bound under tempo scaling, while coda identity remains more stable. Rhythm-only baselines captured lower-tier structure but missed upper-tier sequential dependence. This work provides evidence for a duality-of-patterning-like architecture with a rhythmic lower tier and a null-controlled framework for testing combinatorial structure in acoustic token systems.

Key takeaway

For research scientists developing bioacoustic analysis tools, this study suggests you should incorporate null-controlled frameworks to rigorously test for combinatorial structure in animal vocalizations. You can apply the proposed two-tier architecture model, particularly considering rhythmic rather than purely segmental lower-tier units. This approach helps avoid misinterpreting acoustic similarity as symbolic structure, enhancing the robustness of your findings on communication complexity.

Key insights

The study reveals a two-tier, duality-of-patterning-like architecture in sperm whale codas, with rhythmic lower-tier click combinations and upper-tier sequential dependence.

Principles

• Acoustic similarity can imitate symbolic structure.
• Test combinatorial structure with null-controlled frameworks.
• Abstraction gradients can exist across acoustic tiers.

Method

The method involves computational-linguistic structure discovery from continuous audio using frozen audio encoders, held-out structural tests, per-statistic nulls, and acoustic-null recoverability gates.

In practice

• Apply null-controlled frameworks to other animal vocalizations.
• Investigate rhythmic vs. segmental lower tiers in communication.
• Use tempo scaling to identify abstraction gradients.

Topics

• Sperm Whale Codas
• Duality of Patterning
• Computational Linguistics
• Bioacoustics
• Acoustic Structure Discovery
• Animal Communication

Best for: AI Scientist, Research Scientist

Related on AIssential

• Dolph2Vec: Self-Supervised Representations of Dolphin Vocalizations — Dolph2Vec, a species-specific SSL model, effectively uncovers fine-grained dolphin communication patterns using a novel, longitudinal dataset.
• Tonal parsimony in chord-sequence analysis: combining modulation cost and tonal vocabulary — Tonal parsimony optimizes chord sequence analysis by minimizing modulations and distinct tonal centers.
• UniVoice: A Unified Model for Speech and Singing Voice Generation — UniVoice unifies speech and singing generation by factorizing conditioning and using a null melody token for speech.
• AI Trained on Birdsong Can Recognize Whale Calls — Google DeepMind's Perch 2.0, an AI audio model primarily trained on millions of land-based animal recordings including birds, amphibians, insects, and mammals, has shown unexpected strong performance…
• Recurrence-Based Nonlinear Vocal Dynamics as Digital Biomarkers for Depression Detection from Conversational Speech — Depression alters vocal recurrence structure, detectable via nonlinear state-space analysis of conversational speech.
• AP-GRPO: Anchor-Gated Phonetic Alignment with Policy Optimization for Pathological Speech Reconstruction — AP-GRPO reconstructs pathological speech by aligning SLMs using audible anchors and phonetic compatibility, adapting to disease severity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.