Adding a new concept costs more learning than adding a function under an existing concept

What it is

Adding “archive” under an existing “project” usually lets users keep their old model. Introducing “workspace,” “space,” or “collection” alongside “project” requires relearning object boundaries, permissions, naming, and life cycles. Conceptual integrity therefore treats a new concept as a distinct cost, not an ordinary increment equal to any other feature.

Why it happens

A new concept is not just another word. Users must judge when it exists, what it relates to, which actions it allows, how it is displayed, how failure appears, and how it overlaps with or takes precedence over existing concepts; help, permissions, search, notifications, and migration branch accordingly. A function under an existing concept extends known relations inside a familiar frame. Either choice can be correct, but the cost structures differ sharply.

Studying it

Use controlled prototypes to compare a “new concept” version with an “existing concept extension.” Teach different groups one version, then test explanation, prediction, feature discovery, and cross-task transfer. Record moments requiring questions, misclassification, permission misunderstanding, duplicate-entry use, and help queries. Longitudinal observation can show whether the new concept is still used correctly after a week.

Where it stops holding

New concepts should not be avoided at all costs. When an existing concept has accumulated too many exceptions, unclear meaning, or ambiguous risk, a new concept can be more honest; new input technology, market rules, or responsibility models may genuinely require one. The criterion is not minimum count but whether the new concept is stable, definable, covers a class of tasks, and can replace old special rules.

Applying it

• In requirement review, draw both models—mapping to an existing concept and introducing a new one—then list learning, permission, search, help, and migration costs.
• If keeping a new concept, define its name, life cycle, examples, counterexamples, and relations to core objects before shipping features.
• Launch the new concept with a limited role or sandbox, observe interpretation agreement, then broaden it.
• Track support questions about the new concept and the share of users still completing the task with the old concept as evidence for simplification or withdrawal.

Related

• Same group: B2.25.1 Interface structure should match task structure, not organizational or technical implementation structure · B2.25.2 Conceptual integrity means covering all functionality with as few consistent concepts as possible · B2.25.4 Independent design by multiple people without a shared concept model necessarily creates conceptual redundancy and near-equivalent controls
• Nearby: B2.13 Layering and Progressive Disclosure · B2.23 Internal Consistency
• Search terms: concept learning · feature creep · conceptual cost