The wearer's actions are unintelligible to bystanders

What it is

Someone pinches at empty air, nods, talks. People beside them cannot see the menu, and cannot guess where the hand will go next. Bystander intelligibility is the claim that the wearer’s action lacks a shared visible object, so others cannot read it as intent. Unread, they cannot step aside, cannot join the talk, cannot tell work from absence.

Unintelligible is not the social verdict “looks odd.” It is a missing half of the information: the referent of the action exists only inside the headset.

Why it happens

Everyday action is readable because the object sits in public space: a reach for a cup, the cup is seen, the landing can be predicted. Mixed reality leaves the object in a private display while the body still moves in public. Goffman’s “observable action” splits — the body half public, the task half secret. Bystanders receive a gesture without an object, prediction fails, and they step back, interrupt, or walk into the operating volume.

A narrow field of view makes it worse: the wearer already has to turn to see the object, and the bystander has still less to follow. Phone AR can at least be glanced at when the screen faces out; a headset seals the picture by default.

Studying it

Field observation: have non-wearers watch a wearer finish a task, on video or live, and narrate “what they are doing, where the hand goes next.” Accuracy, the distance they will walk into the operating radius, and interruptions are the dependent measures. Koelle and colleagues’ social-acceptability work splits intelligibility from embarrassment; collect “was it read,” not “would I wear one.”

Independent variables: whether the task object has a public stand-in (monitor, projection), movement amplitude, whether bystanders were told a headset is in use. Dependent variables: narration accuracy, hit rate on the next action, distance of spontaneous avoidance.

Informed lab bystanders overestimate understanding — they know there is a virtual menu. Use uninformed observers, or at least do not lead with “this is AR.”

Where it stops holding

When others present also wear and see the same world-locked objects, action becomes readable again. Casting the first-person view to a side screen can buy some intelligibility back, but what they see is a picture, not the space the hand is about to enter. Near-motionless voice-only use is read as “on a call”; the misunderstanding changes kind, it does not vanish. On an industrial floor that already treats “headset on” as “working,” the intelligibility demand drops and clearance is a procedure, not a reading.

Applying it

• In places with bystanders, give the task a public counterpart (a desk screen, a floor cursor) so the landing of the hand can be predicted.
• When someone else must cooperate, name the object in ordinary language (“I’m moving the drawing on the wall”). Do not gesture only.
• For acceptance, ask someone who has never worn the system to watch the live scene and tell what the wearer is doing. Each wrong step is an unintelligible action. Do not only ask the wearer whether they think others understood.

Related

• Same group：N5.06.2 Cameras raise privacy issues for people being captured · N5.06.3 Social acceptability constrains where the device can be used
• Nearby：N5.11 Bystanders and Social Acceptability · N2.13 Physical Safety of Input (Collisions and Bystanders)
• Search terms：bystander intelligibility · illegible action · spectator view