research-document

Composition Science


authors:

  • ChatGPT date: 2026-07-18 project: Composition Science domain: Cross-Disciplinary title: Evidence Review - Attention, Wayfinding, and Externalized Cognition status: research-draft version: 1.0 llm_ingest: true machine_readable: true

Composition Science

Phase 3: Evidence Review – Attention, Wayfinding, and Externalized Cognition

Executive Summary

This phase tested early hypotheses against findings from cognitive psychology, vision science, neuroscience, architecture, human-computer interaction, and wayfinding research.

The evidence suggests that effective composition is not simply about attracting attention. Instead, it reduces the amount of cognition users must perform internally by externalizing structure into the environment.

This shifts the project from searching for ideal proportions toward modeling the relationship between environmental structure and human cognitive capabilities.


Key Evidence

Attention is a Priority System

Attention is influenced by multiple interacting factors:

  • Physical salience
  • Task relevance
  • Learned expectations
  • Reward history
  • Selection history
  • Action opportunities

Observation

Visual salience alone does not determine what users attend to.

Interpretation

Composition must align perceptual importance with task importance.


Hierarchy Alignment

Four hierarchies should align:

  1. Intended hierarchy
  2. Perceptual hierarchy
  3. Behavioral hierarchy
  4. Task hierarchy

Misalignment creates confusion even when layouts appear visually attractive.


Strategic Visibility

Visibility reduces:

  • uncertainty
  • memory requirements
  • inferential reasoning

Visible destinations and constraints improve navigation and decision quality.


Externalized Memory

Strong compositions keep important information visible instead of requiring users to remember it.

Examples:

  • breadcrumbs
  • progress indicators
  • persistent navigation
  • visible system status
  • contextual cues

This principle appears in architecture, software, maps, dashboards, and industrial design.


Decision-Point Landmarks

Landmarks are effective when they are:

  • distinctive
  • stable
  • visible
  • associated with meaningful decisions

Distinctiveness alone is insufficient.


Salience Competition

Increasing the prominence of one element changes the relative importance of every other element.

Salience is relational rather than absolute.


Mental Models

Humans build simplified internal models that support prediction and action rather than exact copies of environments.

Consistent local cues reinforce robust global mental models.


Learned Attention

Some attention patterns are biological while others arise through repeated experience and convention.

Composition should distinguish innate tendencies from learned expectations.


Candidate Cross-Domain Laws

ID Law Confidence
LAW-COMP-001 Attention Priority Law High
LAW-COMP-002 Decision-Point Landmark Law High
LAW-COMP-003 Coherent Model Law High
LAW-COMP-004 Hierarchy Alignment Law Medium-High
LAW-COMP-005 Externalized Memory Law High
LAW-COMP-006 Salience Competition Law High
LAW-COMP-007 Strategic Visibility Law High
LAW-COMP-008 Local Relevance Law Medium-High
LAW-COMP-009 Learned Priority Law Medium-High

Emerging Theory

Externalized Cognition Hypothesis

Effective composition externalizes cognitive structure.

Instead of requiring users to:

  • remember
  • search
  • compare
  • infer
  • predict
  • orient

the composition itself carries part of that burden.

This hypothesis explains recurring patterns across architecture, typography, cartography, software interfaces, photography, film, and product design.


Recommended Next Research

The next evidence review should focus on Gestalt psychology and perceptual organization.

Research topics include:

  • proximity
  • similarity
  • continuity
  • closure
  • common fate
  • connectedness
  • figure-ground
  • grouping

These mechanisms appear to be the bridge between low-level perception and the higher-level cognitive laws identified in this phase.


Open Questions

  • Which Gestalt principles are biologically universal?
  • Can grouping strength be quantified?
  • How does grouping influence working memory?
  • Can cognitive load be predicted from composition?
  • Which laws are invariant across media?

Current Direction

The project is evolving from a catalog of design advice into a unified, evidence-backed theory describing how environments can be structured to match human perception, cognition, and action.