Salience Ladder Review¶
Audit procedure — instantiates Figure-Ground Structuring
Ranks a field's elements by how strongly each actually pulls the eye, then lays that order against the intended task priority to catch what wrongly dominates.
A design can be individually defensible in every part and still send the eye to the wrong place. Salience Ladder Review is the audit that measures where attention actually lands. It ranks every element in a field by its perceived dominance — using a proxy for pre-attentive pull rather than the designer's intent — and lays that observed "ladder" beside the intended task-priority order. Every element that ranks higher than the task wants it to is a competing-figure defect. Its defining move is that it treats perceived salience as something to measure, not assume: the promo banner that outranks the account balance is caught because the review reads the screen the way a first-time eye does, then also verifies the intended figure clears the minimum contrast a low-vision user needs.
Example¶
A team ships the home screen of a mobile banking app. The intended priority, taken from the foreground-background map, is: (1) account balance, (2) the "Pay" button, (3) recent transactions, (4) a promotional banner. The review squints the screen out of focus — and separately runs a greyscale blur — and ranks what still pops. The observed ladder comes back inverted at the top: the promo banner (a bright saturated gradient) dominates, the "Pay" button is second, and the balance — thin grey type on white — is only third. The audit flags the top two rungs as mismatched: the field is shouting an ad over the number people opened the app to see.
It then checks the accessibility contrast floor and finds the balance text sits around 2.9:1 against its background, below the 4.5:1 minimum, so users with low vision lose the figure entirely. The fixes follow directly: drop the promo's saturation, raise the balance's weight and contrast above the floor. A re-run shows the observed ladder now matches intent and the figure clears the floor for everyone.
How it works¶
The distinguishing method is empirical ranking against declared intent:
- Enumerate and rank by perceived pull. List the field's elements and order them by bottom-up salience, estimated with a proxy — squinting, greyscale, blur, or a computational saliency map — so the ranking reflects what grabs the eye, not what the designer meant.
- Align against task priority. Place the observed ladder next to the intended priority order; any element ranked above its intended rung is a competing figure to demote.
- Check the floor separately. Verify the intended figure's contrast clears the accessibility minimum, so the figure survives for users who depend on redundant or high-contrast cues.
It diagnoses and prescribes direction (raise this, dim that); it does not itself define the figure, delete ground detail, or balance channels over time.
Tuning parameters¶
- Proxy fidelity — squint test versus computational saliency versus eye-tracking. Higher fidelity catches subtler mismatches but costs setup and time.
- Floor threshold — which contrast standard applies (e.g., stricter for body text than large text). A higher floor is safer for impaired users but constrains subtle palettes.
- Ranking depth — just the top few rungs versus the full field. Full ranking is thorough but noisy on near-ties.
- Redundancy requirement — whether the figure must be marked by more than one cue (contrast and size). Redundancy is robust but spends design budget.
When it helps, and when it misleads¶
Its strength is closing the gap between intended and perceived salience — the exact failure where every element is defensible yet the wrong thing pops. The squint test is the folk version of this discipline, and formalizing it catches inversions before users do.[n1]
Its failure mode is that a salience proxy measures bottom-up pull and misses semantic pull: a human face, a red word like "error," or a reader's own name draws attention regardless of contrast, so a clean ladder can still mislead. The classic misuse is cranking the figure's contrast until it tops the ladder — sailing past the point where it manipulates, fatigues, or drowns the ground. The guarding discipline is to rank against task priority rather than maximal pop, and to treat the accessibility floor as a floor to clear, not a target to hug.
How it implements the components¶
Salience Ladder Review realizes the archetype's contrast-tuning machinery:
contrast_salience_gradient— it reads and prescribes the gradient, ranking perceived dominance and saying where contrast must rise or fall so the figure sits at the top rung.accessibility_contrast_floor— it explicitly checks the figure's contrast against the minimum needed by users relying on redundant cues.
It does not implement support_channel_hierarchy — that is Lead-Support Mix Check, its nearest twin: the ladder audits the whole field's perceived dominance order, whereas the mix-check balances one named lead against its supporting channels over time. It also does not define the figure it ranks against — foreground_candidate_definition belongs to Foreground-Background Map.
Related¶
- Instantiates: Figure-Ground Structuring — it verifies the rendered field actually elevates the intended figure.
- Consumes: Foreground-Background Map supplies the intended-priority order the observed ladder is checked against.
- Sibling mechanisms: Foreground-Background Map · Background Load Sweep · Figure-Ground Reversal Test · Baseline Context Band · Lead-Support Mix Check
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Salience Ladder Review operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it ranks a field's elements by how strongly each actually pulls the eye, then lays that order against the intended task priority to catch what wrongly dominates.
Independent corroboration: The frozen evidence defines Salience Ladder Review as 'Ranks a field's elements by how strongly each actually pulls the eye, then lays that order against the intended task priority to catch what wrongly dominates', so its operative form is Assessment, Review & Assurance.
Nearest alternative: Analysis, Modeling & Optimization — Salience Ladder Review includes features of an analytical, modeling, inference, comparison, or optimization procedure that derives insight or a solution, but its defining operation is a bounded evaluation of existing evidence or work that produces a finding or disposition.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Art & Aesthetics
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Ranking visual elements by eye-pull against intended hierarchy is rooted in visual composition and critique.
Related originating lineages:
- Human-Computer Interaction — Human-computer interaction and interface design supplies a parallel or contributing lineage for the mechanism's defining operation: ranks a field's elements by how strongly each actually pulls the eye, then lays that order against the intended task priority to catch what wrongly dominates.
- Psychology — Attention research independently explains perceptual capture.
Review resolution: Both blind reviewers agree that art_aesthetics is the primary historical origin. Explicit reconciliation of alternate_origin_disagreement starts from reviewer_a's mechanism-specific evidence: Ranking visual elements by eye-pull against intended hierarchy is rooted in visual composition and critique. Reviewer A proposed alternates=human_computer_interaction, psychology, origin_mode=convergent, domain_reach=multi_domain, and encyclopedia_synthesis=true; reviewer B proposed alternates=human_computer_interaction, origin_mode=convergent, domain_reach=multi_domain, and encyclopedia_synthesis=true. The final record retains every independently supported alternate from either review (human_computer_interaction, psychology) without an arbitrary cap, selects origin_mode=convergent to represent the combined lineage evidence, and records domain_reach=multi_domain and encyclopedia_synthesis=true. Present-day transfer is recorded as reach and is not treated as proof of historical origin.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] The "squint test" is a working designer's heuristic: blur your vision (or the screen) until detail dissolves, and see which elements still stand out. What survives the blur is what pre-attentively dominates. The ladder review generalizes this into a ranked comparison against intended priority. ↩