Composition Science Research Library

Version: 0.1 (Working Draft)

Purpose

This repository captures evidence, observations, hypotheses, and candidate laws related to visual composition across disciplines including UI design, architecture, editorial design, psychology, neuroscience, typography, industrial design, and art.

Rule #1

No principle without evidence.

Rule #2

Separate observation from interpretation.

Rule #3

Every proposed law must make testable predictions.

Candidate Laws (Hypotheses)

L-001: Law of Dominance

Hypothesis

Every composition has a dominant element, whether intentional or not.

Predictions

• Users should identify the dominant element quickly.
• Multiple competing dominant elements should increase search time and reduce confidence.

Evidence Status

🟡 Early evidence

L-002: Law of Relative Separation

Hypothesis

Perceived grouping depends more strongly on relative spacing than absolute spacing.

Predictions

• Increasing the ratio between groups while leaving internal spacing constant should improve grouping.
• Simply increasing all spacing equally should produce less improvement.

Evidence Status

🟡 Early evidence

L-003: Law of Limited Attention

Hypothesis

Humans perceive only a small region of a composition in high detail at any moment.

Implications

Hierarchy is required.

Everything cannot be equally important.

Evidence Status

🟢 Strong support expected from vision science.

L-004: Law of Visual Economy

Hypothesis

Every visual element carries a cognitive cost.

Predictions

Removing unnecessary visual elements should reduce cognitive effort unless they also remove useful structure.

L-005: Law of Hierarchical Progression

Hypothesis

Good compositions produce a predictable order of discovery.

Evidence Registry

OBS-0001

Observation

Large uninterrupted whitespace often reduces perceived complexity.

Evidence Sources

• Apple product pages
• Editorial magazine layouts
• Minimalist product packaging

Counterexamples

• Trading terminals
• Air traffic control
• Medical monitoring

Confidence

★★☆☆☆

OBS-0002

Observation

Relative spacing appears to communicate grouping more effectively than absolute spacing.

Supporting Disciplines

• Gestalt psychology
• Interface design
• Editorial layout

Confidence

★★☆☆☆

OBS-0003

Observation

Visual hierarchy consistently appears before semantic understanding.

Users typically know where to look before they understand what they are seeing.

Confidence

★★★☆☆

Research Backlog

Vision Science

• Visual acuity
• Crowding
• Peripheral vision
• Contrast sensitivity
• Eye tracking

Cognitive Psychology

• Working memory
• Chunking
• Attention
• Visual search
• Decision making

Gestalt Psychology

• Proximity
• Similarity
• Common region
• Connectedness
• Continuity

Editorial Design

• Reading rhythm
• Headline hierarchy
• White space
• Information pacing

Architecture

• Circulation
• Compression and expansion
• Sight lines
• Wayfinding

Industrial Design

• Affordances
• Visual balance
• Material perception

Research Extraction Template

Citation

Authors:

Year:

Title:

Discipline:

DOI/URL:

Experiment

Research Question:

Participants:

Independent Variables:

Dependent Variables:

Task:

Viewing Distance:

Display:

Results

Measurements:

Effect Size:

Statistical Significance:

Interpretation

Authors' Conclusion:

Our Interpretation:

Composition Implications:

Limitations:

Notes

This document is intentionally conservative.

Every statement should become stronger only as evidence accumulates.

The long-term goal is to derive evidence-based composition principles that can inform style guides, layout generation, component libraries, and AI-assisted interface design.