The classic capacity number came from an absolute-judgment task on a single dimension, structurally unlike most interface scenarios

What it is

The widely quoted "seven" traces back to Miller's 1956 paper, which actually reported two structurally unrelated lines of research that happened to converge on a similar figure. One is absolute judgment: a single stimulus varying along just one dimension (a given pitch, a given loudness) is presented in isolation, and the participant must immediately name which category it belongs to, with no other stimulus present for comparison. The other is memory span: a sequence of items is presented, and the participant must recall them in order afterward. Nothing about these two tasks connects them cognitively — they merely landed near the same number, and the paper itself was built around drawing attention to that coincidence. When later readers remember only the number "seven" and forget it actually came from two separate lines of work, anyone citing it tends to unknowingly smuggle in the absolute-judgment finding too — and that finding has nothing to do with "how many things can be held in mind or seen at once."

Why it happens

An absolute-judgment task measures something specific: when a stimulus varies along a single sensory dimension and there is no other stimulus available for comparison at the moment of judgment, how many categories can a person reliably distinguish and consistently label? This is essentially a ceiling on single-channel information transmission, and it has nothing to do with whether content needs to be held in mind — the stimulus is right there in front of the person, and all they have to do is attach a label to it. This differs structurally from most interface scenarios that actually need a limit: menu options, palette colors, chart categories are typically several options presented at once for comparison and choice, not a single isolated stimulus requiring immediate categorization. Citing "seven" to cap this kind of multi-option, simultaneously-presented scenario implicitly assumes two structurally different tasks share one ceiling — an assumption that has no grounding of its own.

Studying it

To determine which line of Miller's work a cited "seven" actually traces to, go back to the specific task description: if the task presents a single stimulus in isolation and demands immediate categorization with no sequential memory component, it corresponds to absolute judgment; if the task presents a sequence of items and asks for ordered recall afterward, it corresponds to memory span. The two numbers being close is purely coincidental — that coincidence is no license to apply one line's finding to the other's task, and even less to a third kind of scenario that is neither (such as comparing several simultaneously visible options).

Where it stops holding

The absolute-judgment finding isn't entirely without application in interface design — if a control is itself genuinely single-dimensional and requires the user to identify which level it's currently at without any side-by-side comparison (a dial distinguishable only by tactile detents, a judgment of depth on a single continuous color ramp), the task structure there genuinely resembles absolute judgment, and citing that line's capacity number would have a real correspondence. But such scenarios are the minority in interface design; most situations that call for some kind of limit are comparison tasks with several options visible at once, which don't meet this precondition.

Related

• Same group: A6.10.1 Classic memory-capacity numbers don't apply to counting visible options · A6.10.2 Menu-item limits come from search cost, not memory · A6.10.3 Citing a capacity finding requires checking the original task type · A6.10.5 Popular science flattened the number into a universal design rule, detached from its original conditions · A6.10.6 Design limits on password length or color count that borrow this number lack empirical support · A6.10.7 Deciding whether a design limit should invoke memory capacity requires first confirming the user is really doing unprompted recall
• Nearby: A6.02 Working memory capacity · A1.05 Color vision and opponent-channel mechanisms
• Search terms: absolute judgment · Miller 1956 · channel capacity · unidimensional stimulus