Parameter controls are suited to precise adjustment

What it is

“A bit more formal” names a direction, not a stop. Dragging a slider from 3 to 4 is a stop that is shared, repeatable, and tellable to someone else. Parametric precision control means that quantities which already exist — length, temperature, intensity, an enum — are more stably moved with a graduated control than approximated in another sentence. Language is good at stating a goal. Ticks are good at nudging a goal that is already lined up.

Precision here is not “smarter.” It is repeatable action on one dimension. Being able to return to the same tick a second time is what makes it precise.

Why it happens

Once a quantity is named, people’s adjustment strategy becomes climbing a dimension: a little more, a little less, compare to just now. A control externalizes the dimension as a spatial position; step size is set by the scale; two actions can refer to the same state. Language encodes degree as vague predicates (“slightly,” “very”); each decode re-estimates a point; two “slightly”s are not the same point.

There is also communication. In collaboration, “set formality to 4” is an executable instruction; “a bit more formal” lands on a different 4 in each head. A control is a shared metric. Degree words in language are private metrics.

Studying it

The task is already aligned (the letter is already the right object and situation); the only demand is targeted adjustment on a single quantity (120±10 words, formality to a named band). One group iterates with degree adverbs, one group uses a slider or stepper. Dependent variables: rounds to target, overshoot count, agreement when two people independently adjust to the same target. Independent variables: whether the quantity is visible to the user, whether the scale shows numbers, whether the language group can see the last numeric value (even if they cannot edit it directly).

Do not mix in “was the final writing good.” Measure efficiency and repeatability of climbing a known quantity.

Where it stops holding

When the quantity has not been invented yet, there is no dimension for a control to climb, and precision does not apply. Forcing a continuous aesthetic quantity (“more cinematic”) onto a slider fakes precision. In accessibility, fine dragging can be harder than saying another sentence; a stepper or numeric field is then the usable form of precision. Some quantities in images (seed, sampling steps) are not understandable quantities for non-specialists; giving them a control is giving them noise. This entry does not treat whether a control can be reverted, nor how a range displays a capability boundary.

Applying it

• Make quantities the product already treats as stable into controls: length, formality bands, format enums, explicit sampling parameters such as temperature. Do not force people to chase a word count with “a bit shorter.”
• Show the current value. Adjustment is relative to now, not a re-estimate from language each time.
• Match step size to decision grain: numbers or steppers for word count, three to five bands for tone. Do not give tone a fake-precise 0–100 slider.
• Check: two people apart, same draft, target “about one hundred twenty words, formal band.” Under a control path their endpoints should land near the same tick; if a language-only path scatters, precise adjustment was not offered. Then watch overshoot: whether the language group ping-pongs “too short / too long” several times.

Related

• Same group: L2.04.1 Natural language is suited to expressing intent · L2.04.3 Having both beats choosing one · L2.04.4 Natural language is strong at goals and constraints; parameter controls are strong at degree and numeric values · L2.04.5 Parameter controls provide a revertible determinate state; natural language does not · L2.04.6 When both exist, which one overrides the other must be explicit, or conflict is unpredictable · L2.04.7 Filling parsed language back into controls lets users see what the system understood · L2.04.8 A control’s value range is itself a display of the capability boundary
• Nearby: L2.06 Granularity of Control · L1.03 Visualizing Uncertainty · L2.05 Iterative Revision
• Search terms: parametric precision control · magnitude adjustment · shared scale