Arbitrary mappings need labels or memory support

What it is

An arbitrary mapping has no control–result relation that can be directly inferred from spatial arrangement, physical causality, or familiar convention—for example an abstract icon, an unlabeled keyboard shortcut, or color coding unrelated to the result. Such mappings are not unusable, but people need labels, examples, persistently visible reference, or long-term memory to know how to act.

Why it happens

When an action–outcome relation cannot be derived from the current situation, people must retain its rule in working memory or retrieve it from an external cue. First use brings pausing, searching, and guessing; learned correspondences also decay after disuse. When multiple arbitrary mappings are similar but differ, interference grows and a person may transfer one rule incorrectly to another place.

Studying it

Test first use, delayed return, and switching among functions separately. Measure whether people can predict a correct action without instruction, find needed reference, and recover from error. Asking them to explain an icon, shortcut, or code distinguishes genuine understanding from accidental position memory. Compare performance with and without labels, tooltips, command search, and examples to estimate the value of external support.

Where it stops holding

Arbitrary does not mean bad: abstract symbols can save space and speed frequent expert action, while shared platform or professional conventions may already distribute learning cost through history. The problem is turning unfamiliar, low-frequency, or high-consequence actions into unsupported memory tests. Excessively long labels also add scanning cost, so support should match task importance.

Applying it

• Provide stable labels, tooltips, command search, or adjacent examples for controls that cannot be inferred at a glance, so their rule can be read externally.
• Keep the meanings of symbols, shortcuts, and codes consistent across the system to reduce interference among similar rules.
• Prefer visible explanation and confirmation for infrequent or high-consequence actions; do not assume one tutorial creates permanent memory.

Related

• Same group: B2.04.1 The control–result relationship determines learning cost · B2.04.2 Spatial layout correspondence is the strongest natural mapping · B2.04.3 Cultural-convention mappings cannot be assumed across cultures
• Nearby: B2.03 Signifiers · B2.07 Consistency
• Search terms: arbitrary mapping · memory load · external memory