## Cognitive Landscapes  
### Visualizing Computation, Criticality, and Path Emergence in Intelligent Landscape Design

**Abstract**

This paper introduces *Cognitive Landscapes* as a framework for intelligent landscape design in which land, infrastructure, social systems, and communities are understood as a single coupled cognitive system. Rather than treating landscapes as static forms or purely representational objects, we frame them as systems that perceive, act, remember, and compute through intensive gradients, asymmetric interaction rules, and emergent paths.

Building on a seminar-based design pedagogy at Harvard Graduate School of Design, the work uses time-lapse imagery, environmental sensing, AI-assisted pattern inference, and drawing to reveal the invisible drivers of spatial form. Temporal resampling shifts imperceptibly slow or fast processes into perceptible regimes, allowing students to identify intensive gradients and rule asymmetries that give rise to extensive form, following DeLanda’s distinction between intensive and extensive space.

Central to this framework is the claim that **paths are not merely outcomes but computational processes**. Least-time and least-action paths emerge through distributed exploration and reinforcement, a universal algorithmic structure visible across rivers, fires, floods, traffic flows, ant systems, plasma channels, and even the transactional interpretation of light, where the photon is understood as a completed path rather than a moving object. These systems operate within a narrow critical regime in which memory and computation are possible; outside this regime, coherent paths cannot form.

The paper extends the concept of landscape sheds to include firesheds, floodsheds, and communication sheds, arguing that information flow is as critical as water or fire flow in community adaptation. Using student analyses of stream meandering, wildfire percolation through housing, post-fire flooding, and civic communication networks, the work demonstrates how landscapes and communities co-compute structure through coupled perception–action loops.

By integrating ecological cognition, path computation, and digital visualization, this research reframes intelligent design as **immanent rather than imposed**. Intelligence is neither external authorship nor post-hoc selection, but an emergent property of systems operating near criticality, where landscapes compute viable paths and communities knowingly participate in that computation. As a contribution to digital landscape architecture, the paper formalizes a reproducible pedagogical and methodological approach for visualizing computation in landscapes and designing the conditions under which adaptive form can continually emerge.