Tal Boger¹, Shari Liu¹, Chaz Firestone¹; ¹Johns Hopkins University

What connects a sharply twisted shape, a many-layered melody, and the multisyllabic string “animipatorun”? These items are unrelated in nearly every aspect; they span different modalities, arise from different domains, and have independent properties. Nevertheless, they seem unified by their *complexity*: Each is informationally dense relative to prototypical stimuli of its kind (cf., a square, major scale, or short string). Does the mind appreciate the complexity these stimuli share, even across dramatically different properties? Here, 4 experiments demonstrate *transfer* across these different stimuli, suggesting that a ‘universal’ representation of complexity exists in the mind. In Experiment 1, participants learned a reward rule for simple and complex shapes; selecting a complex shape was worth more (or less) points than selecting a simple shape. After this learning phase, participants saw new stimuli that also differed in their complexity: two arrays of colored dots, one uniform and the other highly varied. Without any further instruction, subjects transferred the reward rule to the dots, spontaneously selecting the more complex (or simpler) dot array. Experiment 2 generalized this pattern to audition: Subjects who learned that complex shapes were worth more points spontaneously selected complex melodies. Experiment 3 extended this result even further, finding successful transfer from shapes to letter-strings. In each case, this transfer arose bidirectionally. Finally, Experiment 4 tested the automaticity of such transfer. In a Stroop-like task, two shapes of differing complexity appeared above two letter-strings of differing complexity, and participants judged which shape (or letter-string) was more complex. Though only one stimulus class (either the shapes or the letter-strings) was task-relevant, participants were faster to judge the complexity of the target stimulus when the task-irrelevant stimulus was congruent in complexity. We suggest that visual, auditory, and linguistic complexity are ‘unified’ in the mind, supporting spontaneous and automatic transfer across modalities.

Acknowledgements: NSF BCS 2021053, NSF GRFP