Do Humans (and Animals) Think in Poetic Language?

The search for thought’s representational format has traditionally focused on Mentalese, neural codes, predictive processing, embodied sensorimotor models. Here we examine another hypothesis: human cognition may operate by default through a structurally poetic mode characterized by metaphorical mapping, rhythmic/repetitive patterning, imagistic compression, controlled ambiguity, acoustic structuring and affective resonance. These features correspond to identifiable computational and neurobiological mechanisms. Metaphor and analogy recruit fronto-parietal networks for cross-domain alignment, enabling dimensionality reduction and transfer across abstract spaces. Rhythmic and metrical structure entrain delta–theta–gamma hierarchies, improving prediction and memory via phase-aligned hippocampo-cortical loops. Imagery activates hippocampal–visual circuits and default-mode subsystems, supporting high-level generative simulation. Ambiguity/polysemy sustains parallel semantic states in frontal networks, enabling probabilistic inference under uncertainty. Sound patterning exploits phonological memory and auditory statistical learning, enhancing signal-to-noise resilience. Emotional resonance engages limbic and reward pathways, modulating salience and long-term retention. Comparative evidence indicates that structurally homologous “proto-poetic” systems exist across species. Birdsong and whale song show hierarchical, rhythmic sequence generation with cultural drift; honeybee dances encode spatial data cross-modally; primate calls display affix-like combinatoriality and contextual ambiguity; frog and insect chorusing reflects large-scale oscillator coupling; and ritual displays compress identity and fitness cues into symbolic form. These convergent data support the view that poetic structuring constitutes an evolutionarily conserved cognitive technology optimized for memory efficiency, predictive precision, adaptive flexibility and social synchrony. Reframing poetry as a computational architecture of thought motivates new empirical programs integrating neuroimaging, comparative biology and formal models of information compression, oscillatory coordination and cross-domain mapping.