Without boundary feedback, users misread out-of-bounds as a system fault or a wrong gesture

What it is

If crossing the boundary has no dedicated account, users must invent a cause from “no response.” Two common wrong attributions are a system fault (frozen, camera dead, network down) and a wrong gesture (mine does not look like the tutorial). Both skip the stance that actually needs changing. Once the misread sticks, the next behaviors are rebooting the device or making the motion more exaggerated, which steps on the boundary harder.

Why it happens

Attribution draws from the hypothesis set that is available. Visitors at a public screen already carry “electronics break” and “I may not know this vocabulary”; they rarely carry “there is a software box in the air.” With no boundary signal, the box hypothesis never enters the set. People who pick “fault” hunt for status lights, call staff, or quit. People who pick “wrong gesture” match the diagram, enlarge, slow down, switch hands—repairs that at body scale push joints further out or add self-occlusion. Teaching materials that demonstrate only hand shape, never stance, raise the prior on “wrong gesture” further. The misread is not merely a mood. It drives strategies that make recovery harder.

Studying it

Induce out-of-bounds with no boundary feedback, then run a structured debrief: participants rank causes from a list that includes fault, gesture, stance, not activated, and other, and demonstrate what they would do next. Repeat with boundary feedback present. Compare the distribution of the first cause and whether strategy points at stance. Coders should split “I will try the same motion again” from “I will change place.” Staff interventions in the venue are an ecological index of misread: a system treated as faulty consumes operations.

Where it stops holding

When someone beside them says “stand a bit forward,” social information corrects the misread and lab attribution distributions look optimistic. On equipment used daily by staff, the fault hypothesis drops; the gesture hypothesis may remain. Drawing a tracking skeleton on an out-of-volume hand pushes attribution away from “the camera is dead” without automatically pointing at stance—it may be rewritten as “the algorithm is broken.” Children use the category “system fault” less and simply leave; the misread then looks like abandonment rather than a wrong repair.

Applying it

• When out of volume, name stance as the cause; do not leave users to guess between fault and gesture. Do not put “unrecognized” in the same sentence.
• Teach hand shape and foot position together, lowering the prior that “it can only be the gesture.”
• Count on-site “reported broken, camera fine” and “user made the motion larger and larger.” Both are typical products of missing boundary feedback and should be billed separately from true faults and true recognition errors.

Related

• Same group: C4.22.1 The boundary itself is invisible; users perceive it only through whether actions take effect · C4.22.2 Boundary crossing needs its own feedback, distinct from recognition-failure feedback · C4.22.4 Actions near the boundary should receive progressive cues rather than failing abruptly at the edge
• Adjacent: C4.20 Interaction volume versus capture range · C4.05 Three dimensions of gesture evaluation
• Search: misattribution · out-of-bounds · fault versus gesture