Skip to content

Theoretical Foundations

The Problem with "Place"

Traditional GIS treats space as a container with objects located within it. But place is not simply location — it emerges from:

  • Repeated encounters between agents and spatial extents
  • Accumulated familiarity that grows and decays
  • Affect — the emotional character of experience
  • Meaning — symbolic interpretations attached to locations
  • Practices — routines and habits that pattern our engagement

Core Premise

Place is not a primitive. Place emerges.

Chora models place as an emergent subgraph from a typed, temporal, heterogeneous graph of encounters, rather than as a predefined spatial category.

Design Principles

Principle Implementation
Relational Primacy Platial qualities on edges, not nodes
Encounter-Centricity Encounters as first-class objects
Epistemic Separation OBSERVED → DERIVED → INTERPRETED
Probabilistic Representation Uncertainty throughout
Temporal Explicitness All entities have lifetimes
Theory-Encoded Computation Derivations embody platial theory

The Platial Graph

Agent ──PARTICIPATES_IN──► Encounter ──OCCURS_AT──► SpatialExtent
                              │
                              ├──HAS_CONTEXT──► Context
                              ├──EXPRESSES──► Affect
                              ├──REINFORCES──► Familiarity
                              └──BELONGS_TO──► Practice

Epistemic Levels

All data is explicitly categorised:

  • OBSERVED — Direct measurements (GPS traces, check-ins)
  • DERIVED — Computed from observations (familiarity scores)
  • INTERPRETED — Semantic/symbolic meanings

This separation ensures that uncertainty and provenance are preserved through all transformations.

Familiarity Dynamics

Familiarity is modelled with: - Reinforcement on each encounter (saturating growth) - Decay over time without encounters (exponential)

# After 5 visits
familiarity = 0.16

# After 14 days without visits
familiarity = 0.08  # Decayed by half

Place Emergence

A "place" in Chora is not a stored entity but a computed view:

place = extract_place(graph, extent_id, agent_id)
# Returns: EmergentPlace with familiarity, affect, meanings

This allows multiple agents to have different "places" at the same location.