A continuous stream of sensory information reaches our senses, yet we perceive discrete, meaningful chunks. In language, for example, we perceive words, phrases, and sentences even though no physical signals mark their boundaries. How are these boundaries formed, and how do they influence subjective perception?
One possible consequence of segmentation is the distortion of subjective time perception. Familiar languages may feel slower, while unfamiliar ones often seem faster. Research suggests that crossing event boundaries can compress or stretch time perception, but the exact nature of this distortion and its interaction with hierarchical segmentation levels—such as syllables, part-words, and words—remain unclear.
To address these questions, we combined a statistical learning (SL) task with a pause adjustment (PA) task. In the PA task, participants adjusted pauses between syllables in structured and random streams until all pauses were perceived as equal in length, minimizing memory biases.
Results showed significantly greater pause distortions in structured streams compared to random streams. Within structured streams, distortions were larger for low transition probability (TP) pairs than for high TP pairs. Linear regression analysis revealed a significant correlation between distortion differences in TP pairs during the PA task and reaction time differences in the SL task, suggesting a link between segmentation ability and time perception distortion.
These findings demonstrate that event segmentation distorts subjective time perception and highlight how this distortion varies across hierarchical levels of segmentation.