Older yet Faster: Representational Development Explains Accelerated Semantic Processing in Older Adults

Slower responses in older age are often attributed to cognitive decline. This view overlooks how lifelong learning reshapes mental representations. Since learning continuously updates these representations, behavioral changes in older adults may therefore reflect learning, not decline. Across three studies using semantic decision tasks, we found that older adults consistently outperformed younger adults in processing speed, as estimated by drift rate in diffusion decision modeling. We show this surprising advantage contradicts the prevailing decline-based accounts and instead supports a learning-based framework: Over time, older adults’ semantic representations become more distinct and better aligned with environmental regularities, enabling more efficient processing when task demands match accumulated experience. We validated this account through semantic rating tasks and learning models using inputs from natural language. Crucially, representational development also predicts that older adults’ processing advantage should diminish when learning is disrupted. We tested this by identifying a natural case of disrupted learning: words whose meanings shifted between 1970 and 2000, such that only older adults experienced the change. As predicted, semantic change reduced older adults’ advantage in both processing speed and representational distinctiveness. Our findings challenge the assumption that aging inevitably impairs cognition. Instead, we demonstrate that age-related behavioral changes reflect the dynamic interplay between prior knowledge and task demands, advancing a more balanced view of cognitive aging grounded in lifelong learning and representational development.