In one paragraph
Styled paragraph elements must not be used as headings addresses visual section titles that lack equivalent programmatic heading structure. Not every large or bold paragraph is a heading, and appropriate ARIA semantics can also matter in a real implementation. Review its actual function and prefer native headings where suitable. The supplied experimental rule does not run in the current report, so manual outline review is required.
Why it matters
People can recognise a section visually from size, weight, and spacing. Assistive navigation needs the corresponding structure communicated programmatically. A plain styled paragraph used as a section title may therefore be absent from heading lists or level-based navigation.
WCAG's information-and-relationships and heading requirements concern meaning, not banning all bold paragraphs. Introductory emphasis, pull quotes, and prices do not automatically become headings because they are prominent. Do not convert every typographic accent into another outline entry.
Native heading levels can express relationships while CSS controls appearance. A component library should not force authors to choose between correct semantics and the desired visual design. A valid alternative implementation still needs appropriate role, level, and real structure rather than a mere style guess.
How WebAuditOS evaluates this finding
The supplied p-as-heading description uses paragraph content and style comparisons with a following paragraph, including length, punctuation, relative size, and weight heuristics. Those cannot decide the author's actual content structure reliably in every layout.
The source states this experimental rule is not selected by the current WCAG-tag run and therefore creates no finding. Its nominal serious impact is not a reported severity. The custom heading-order rule only examines headings that already exist and cannot independently discover every styled paragraph title.
How to fix it
- Inspect the reader's actual section structure. Compare visual boundaries, content meaning, and the accessibility-tree outline, not only computed font sizes.
- Replace a true visual section title with an appropriate native heading when practical. For example:
<h2 class="section-title">Payment methods</h2>
<p>Available methods depend on your order and location.</p>This is illustrative content. Choose the correct level for the real page rather than treating h2 as the only permitted replacement.
- Keep typography in maintainable styles so the semantic element can use the intended appearance. A tag change need not adopt every browser default margin or font size.
- Fix the shared CMS block or design-system component producing false headings. Offer authors clear semantic choices instead of only large, medium, and bold visual controls.
- Review any existing role=heading and aria-level solution in its actual host context. Prefer straightforward native structure where possible, but do not claim a paragraph's visual tag alone disproves all programmatic semantics.
- Test the complete hierarchy and meaningful labels. Do not insert empty heading levels solely to satisfy a sequence checker, or turn every short standalone message into a heading.
- Check responsive states, duplicated mobile titles, translated text, and dynamically loaded sections. The initial source and a single screenshot may show different structures.
- Record manual review and rescan related heading checks. A clean report cannot establish that this disabled heuristic inspected every candidate or that every prominent phrase belongs in the outline.
What the fix can break
Changing p to h2 or another level can activate global styles and alter spacing. Preserve the intended design and inspect narrow layouts.
Renamed headings can change generated anchors and table-of-contents links. Keep stable references where users or external links depend on them.
Do not remove useful emphasis from ordinary prose merely because it resembles a heuristic candidate. Repair true structural omissions rather than flattening the visual hierarchy.
Questions
Is bold text in a paragraph prohibited?
No. Emphasis is legitimate; the issue is a real section title whose structure is not appropriately communicated.
Does the heading-order check detect missing semantic headings?
Not generally. It evaluates the headings in its own input, not every visually implied section.
Does no finding mean these paragraphs were reviewed?
No. The supplied configuration does not run this experimental rule.
Which related checks help?
Review heading levels, bypass mechanisms, and the automation limits guide.
What severity does this get in the report?
None, today - the rule is experimental in axe-core and our scan does not run experimental rules, so it never produces a finding. axe declares the rule serious; if it ran and failed, the mapping of critical and serious to high, moderate to medium and minor to low would make it high.