C3.21.3Nonlinear gesture vocabulary loaddesignresearch

Memory load grows nonlinearly with gesture count

Aliases: gesture vocabulary · command-set growth · vocabulary size

What it is

Each extra gesture does not add one unit of load; it interferes with the ones already stored. When the vocabulary grows from four to eight, the new stroke competes for encoding with the old, and confusions rise combinatorially. Elicitation studies such as Wobbrock, Morris, and Wilson keep finding a small consensus set of plausible motions; designers who keep stuffing private strokes into the table should not expect memory to scale linearly.

Why it happens

Each gesture is a stimulus–response link. As the set grows, similar trajectories (two fingers, all downward) share cues, so retrieval activates a cohort that must be inhibited. That is a classic fan effect: the more responses a cue points to, the slower and more error-prone the extraction. Direction × count × duration looks like a large space; the number of discrete commands people can hold stably is small. Continuous manipulations (scroll, pinch, rotate) can share one physical metaphor and do not occupy command slots; the load spike comes from arbitrarily bound shortcuts.

Studying it

Grow the command-gesture set on one UI, holding metaphor strength roughly constant. Measure trials to criterion, delayed confusion matrices, and time. The interesting errors land on neighbors (two-finger pull versus one-finger pull-to-refresh). Set size is the factor; do not mix this with a discovery test in the same session—participants must already know that each gesture exists.

Where it stops holding

Counting pinch, scroll, and rotate inside “gesture total” inflates the vocabulary: they are continuous manipulations, not discrete commands. Reusing a small system set across apps parks the load on the platform, not on one title. Expert CAD users can hold a larger table if they use it daily. Nonlinear growth describes memory, not how many hotspots fit on a screen.

Applying it

  • Cap private discrete gestures in the single digits; overflow goes to menus.
  • A new stroke must sit far from existing ones on a direction × count map; forbid adjacent cells.
  • Accept with a confusion matrix: if adding one gesture raises errors on old ones, delete it rather than teach harder.

Related

  • Same group: C3.21.1 Gestures depend on recall rather than recognition · C3.21.2 How fast infrequent gestures are forgotten
  • Adjacent: C3.11 Two-finger and multi-finger taps · C3.20 Gesture discoverability
  • Search: gesture vocabulary · fan effect · command set size

Cards in the same group

Quick Actions

Share

Share this page

ios_share

https://hci.top/en/handbook/C3.21.3