Direct grab matches everyday experience
Aliases: reach-and-take · isomorphic grab · natural grasping
What it is
A cup sits in front of you. Fingers close. The cup comes along — no ray, no two-step “select then pick up.” Direct grab makes taking a virtual object the same action as taking a real one: the hand arrives at the object’s place, the hand shape closes, the object belongs to the hand. What it matches is a few decades of practice, not a newly taught spatial grammar.
Matching experience is not the same as matching feel. Experience is “I guessed the right move on the first try.” Whether there is resistance, and whether the object is in range, are separate claims.
Why it happens
People store spatial locations in the body. Reaching for a seen object lets vision name the target and proprioception name where the arm already is; the two close at arrival. That loop is used daily on cups, doors, and pages. It does not first translate the object into “a target a ray must hit.” Direct grab puts the virtual object on that same loop: the object has a coordinate the hand can reach, and a closing shape maps to acquisition.
The learning cost is therefore not “understand the metaphor” but calibration. If the virtual hand and the real hand sit a few centimetres apart, experience will close on the real hand first and be corrected by vision. When the offset is small, the correction happens inside one movement and hardly feels like learning. When it is large, experience becomes interference and people start servoing on the visible fake hand — no longer everyday grab, a new visual-servoing skill.
Studying it
Have the same people move a set of objects with direct grab and with indirect distant acquisition (ray pick-up, menu teleport). Compare attempts to first success, not only speed after they have settled.
Independent variables: acquisition method, positional offset between virtual and real hand, whether the object resembles an everyday grabbable (cup, block, meaningless solid). Dependent variables: failures before first success, whether the hand goes to the object’s real location or to a ray landing, verbal “how did you know how to take it.”
The first minute is more informative than the tenth. Experience-match shows up as a zero-training strategy: do people simply put their hand out. After practice both methods can be fast; that measures fluency, not everyday fit.
Where it stops holding
Abstract controls with no graspable analogue — floating text labels, settings with no volume — have no grabbing experience to transfer, so direct grab does not become obvious by itself. When tracking delay leaves the virtual hand visibly lagging, the everyday “arrive and hold” fails to meet vision and people abandon the loop. Children’s grasping often slaps with an open hand rather than pinching; a mapping that only accepts a pinch turns experience away at the door. Treating “matches experience” as “therefore always faster” is wrong: distant objects are often faster by ray. Direct grab’s advantage is near, first-try, in the seconds that should not need a manual.
Applying it
- Near, volumetric objects people would pick up in life should default to reach-and-close. Do not force select-then-button.
- Align virtual and real hands closely enough that the offset can be corrected inside one reach. Do not make people first learn “you grab the visible fake hand.”
- Map closing shapes onto everyday grasps: pad pinch for small things, a wrap for large ones. Recognising only one pinch excludes slap, cradle, and hook skills people already have.
- How to check: sit someone new to the system in front of an in-reach object with no instruction. If the first move is a reach to take, the match holds. If it is waving a ray or hunting a button, acquisition never joined the everyday path.