The Decorative Image Myth Is Costing Real Users Real Access

David
alt textwcag 1 1 1decorative imagesweb accessibilitycompliance audit

David · AI Research Engine

Analytical lens: Balanced

Higher education, transit, historic buildings

AI-assisted · Source-linked · Editorially reviewed · Methodology

Trust note

This article was drafted with AI assistance, reviewed against accessibility.chat editorial standards, and should be treated as research and education rather than legal advice. We prioritize primary sources and correct material errors.

3D illustration of a checklist with checkmarks on a rolled paper sheet against a blue backdrop.
Photo by Bastian Riccardi on Pexels

"We agree that poo is not decorative. Remove the picture and save yourself the hassle."

Adrian Roselli's blunt conclusion to his updated alt text decision tree lands harder than it might first appear. The poo test — would you, as the author, care if your decorative image was swapped for a picture of excrement? — is deliberately absurdist. But the logic underneath it is airtight. If the image carries enough identity, mood, or intent that you'd object to replacing it, then it was never decorative. It was always communicating something. And if it was communicating something, a screen reader user deserved to know what.

Roselli's October 2026 post, "Decorative Images…" (opens in new window), stakes out a deliberately absolutist position: decorative images, as a practical category, don't exist in narrative content. That's a provocative claim. It's also, on examination, largely correct — and the implications for how accessibility practitioners and developers approach WCAG 2.1 Success Criterion 1.1.1 (opens in new window) are worth unpacking carefully.

What WCAG 1.1.1 Actually Permits

WCAG 1.1.1 (Non-text Content) (opens in new window) permits empty alt attributes for purely decorative images — those that convey no information and serve no function. The operative word is purely. The standard doesn't define a broad exemption for images an author intends to be decorative. It defines a narrow exception for images that genuinely contribute nothing to the user's understanding of the content.

The gap between intent and reality is where most implementations fail.

Image TypeWCAG 1.1.1 RequirementCommon MistakeCorrect Approach
Purely decorative (no meaning)alt=""Omitting alt entirelyEmpty alt, not absent alt
Mood/theme imageDescribe mood or contextalt="" or alt="image"Brief contextual description
White space fillerRemove from contentAdding generic alt textFix the layout instead
Complex graphicFull text alternative on pagePartial alt textSupplementary long description
Functional (link/button)Describe destination/actionFilename as altAction-oriented alt text
Image of textReplicate the text in altalt="decorative"Exact text from image

The pattern across these categories is consistent: the decorative exemption gets invoked too broadly, too early, and too conveniently.

The Decision the Author Is Actually Making

For a compliance officer working through an audit, Roselli's framework reframes the question productively. The traditional alt text decision tree asks: is this image decorative? His adapted version asks something harder: can you defend that claim to a user who can't see it?

His two-question test cuts through the rationalization that accumulates around these decisions:

  1. If the image were removed entirely, would any readers care?
  2. If the image were replaced with a picture of poo, would the author care?

A "yes" to either question means the image isn't decorative. The image is doing work — aesthetic work, contextual work, emotional work — that a non-sighted user is being excluded from. That exclusion is the access barrier WCAG 1.1.1 was written to prevent.

This connects to a broader pattern our research has documented: automated testing tools flag missing or empty alt attributes, but they cannot evaluate whether an empty alt is appropriate. As our analysis in Beyond Detection: Why Context Separates Automated Testing from Manual Audits shows, automated tools catch at most 37% of real accessibility barriers. The decorative image question sits squarely in the 63% that requires human judgment.

White Space Is a Design Problem, Not a Content Problem

Roselli makes a point that deserves emphasis: when an image exists to fill white space, that's not a content decision — it's a layout failure. The solution is not an empty alt attribute. The solution is fixing the design.

This matters practically. A compliance officer reviewing a site that uses stock photography to break up text-heavy pages needs to ask whether those images are genuinely content-neutral. Most aren't. Stock photography carries implicit messages about who belongs in a space, what professionalism looks like, what the organization values. These aren't decorative signals. They're representational choices — and screen reader users are being told, implicitly, that those choices don't apply to them.

The Northeast ADA Center (opens in new window) has consistently emphasized that accessibility failures in digital content often reflect organizational assumptions about who the primary audience is. When images are systematically marked decorative without rigorous evaluation, the operational message is that visual experience is the real experience, and non-visual access is a secondary accommodation. That framing inverts the actual Title III obligation (opens in new window) to provide equal access — not equivalent access, not approximate access, but equal.

What WCAG 3 Signals About the Future

Roselli links to WCAG 3 issue #878 (opens in new window), a September 2026 proposal to update the definition of decorative images within the emerging standard. The fact that this definitional question is being actively revisited in WCAG 3 development suggests the accessibility community recognizes that the current framing creates too much interpretive latitude.

For practitioners, this is a signal — not a directive. WCAG 3 is not finalized, and WCAG 2.1 remains the operative standard (opens in new window) for most legal and contractual obligations. But the direction of travel matters. Organizations building accessibility programs now should anticipate that the decorative exemption will narrow, not expand, as standards evolve. Building a culture of rigorous alt text evaluation now is both better practice and better preparation.

This connects to what our research on compliance framework complexity identifies as a recurring organizational failure: teams optimize for the current minimum standard rather than building durable accessibility judgment. When standards shift, those teams scramble. Teams that have internalized why alt text matters — because real people need it to access real content — adapt without crisis.

How to Audit Alt Text Decisions

Roselli's framework translates into concrete audit steps:

Flag every empty alt in your next audit. Don't accept empty alt as a pass. Treat it as a question requiring justification. Can the content owner answer both of Roselli's tests? If not, the image needs a description or needs to be removed.

Separate the iconography question from the narrative image question. Roselli explicitly scopes his argument to narrative content, not icons used in interactive controls. Functional images in buttons and links follow a different decision path — one that requires describing the action or destination, not the visual.

Push back on "white space" as a rationale. When a content owner says an image is decorative because it's just filling space, the correct response is: fix the layout. Don't paper over a design problem with an empty alt attribute.

Document your reasoning. For images where the decorative determination is genuinely defensible, write down why. "This image was marked decorative because it was removed in a content test and no users flagged a change in meaning" is a defensible record. "We thought it was decorative" is not.

Roselli's poo test is funny. The access failure it's designed to prevent is not. Every image marked decorative without rigorous evaluation is a quiet decision that someone else's experience of the content doesn't matter enough to describe. That's worth arguing about — and worth fixing.

About the David lens

A balanced lens that weighs competing considerations before recommending. Applied to higher education, transit, and historic-building access questions.

David is an AI analyst lens, not a human staff member. It helps frame this article through a consistent accessibility perspective.

Specialization: Higher education, transit, historic buildings

View all articles using this lens →

Transparency Disclosure

This article was drafted with AI assistance and reviewed against our editorial methodology. We disclose that process so readers can judge the work clearly.