Matterhorn 17-001 · PDF/UA-1 UA1:7.7-1

Equations tagged as text or pictures instead of Formula

WarningCheckpoint 17: Mathematical ExpressionsWCAG 1.1.1 (A)

An equation is tagged as a paragraph or a figure rather than a Formula.

The condition, as the protocol states it. “Content is a mathematical expression but is not tagged with a <Formula> tag.”

Matterhorn Protocol 1.1, checkpoint 17, index 17-001, section UA1:7.7-1. The protocol marks this as requiring human judgement.

What a screen reader actually does

The expression is announced glyph by glyph in visual order — meaningless for anything beyond a simple sum.

What PDF/UA asks

PDF/UA-1 clause 7.7 puts every mathematical expression in a Formula element, with /Alt giving it in words, and applies the Unicode mapping rules to its characters. Formula is one of ISO 32000-1’s illustration types, alongside Figure and Form. Condition 17-001 is the first step failing: an expression that is not in a Formula at all.

/P    "The sum is bounded by the integral:"
/Formula
      /Alt (The sum over i of x sub i is at most the integral of f of x, d x.)
      [the painted expression]

The protocol marks it as a human condition, because only a person can say that a given run of characters is an expression rather than prose that happens to contain a ≤.

What the reader hears

Tagged as a paragraph, an expression is just text. A screen reader reads the characters in content order and names the symbols however its speech engine names them. Everything that was carried by position is gone: a superscript is an ordinary digit set smaller and higher, so x² is read as x 2; a fraction bar is a drawn line, not a character, so a over b is read as a b. For anything beyond a simple sum, the result cannot be reassembled by ear.

Tagged as a figure, the reader announces a graphic and its alt text if it has one. That at least delivers words, but it tells every piece of software that the expression is a picture.

Where it comes from

Equation objects in Office documents. The old Equation Editor objects — retired from Microsoft 365 in 2018 — and MathType objects are embedded objects, and they reach the PDF as pictures. Microsoft now offers Convert to Office Math, and Convert All to Office Math for a whole document, which turns them into native equations. For native Office Math, Microsoft says Word on all platforms, and PowerPoint for Windows from version 2602, now write MathML into the exported PDF. (Microsoft 365 Insider blog, February 2026) Check the result in the Tags panel rather than assuming the export uses Formula.

Auto-tagging. A tagger that does not recognise equations sees text and graphics, and tags an equation as whichever it looks like.

LaTeX. Without the tagging support of recent LaTeX releases, a pdfLaTeX document is not tagged at all; with it, maths handling depends on the setup. See LaTeX to accessible PDF.

How to check it free

Neither Acrobat nor veraPDF tests for this. Acrobat’s Accessibility Checker has no rule for equations in Adobe’s list of its rules, and veraPDF’s PDF/UA-1 rule 7.7-1 tests that a Formula has /Alt or /ActualText — it cannot find an expression that was never tagged as one. (veraPDF PDF/UA-1 rules)

So look. In Acrobat Pro, open the Tags panel (View → Show/Hide → Side Panels → Tags), turn on Highlight Content from its Options menu, and for each equation on the page find the tag that covers it. <Formula> passes this condition. <P>, <Span> or <Figure> does not.

How to fix it free, in Acrobat Pro

Adobe’s Reading Order tool has a Formula button, and Adobe’s own note on it is that speech software may handle formulas differently from text, so alternate text is worth adding. (Adobe: Reading Order tool)

  1. Choose All tools → Prepare for accessibility, then Reading Order in the left pane.
  2. In the dialog, select Show Page Content Groups, then Page Content Order.
  3. Select the expression’s region — drag around it, or click its number.
  4. Click Formula.
  5. Give it /Alt: in the Tags panel, select the new <Formula>, choose Properties from the Options menu, and type the expression in words in Alternate Text on the Tag tab. That is condition 17-002.

Two limits, both from Adobe’s documentation. The Reading Order tool is meant for files Acrobat tagged itself, and Adobe recommends fixing the source where you can. And its regions are rectangles, so an expression in the middle of a line may not be separable with it at all. If the expression already has a tag of its own, change that tag’s type instead: select it in the Tags panel, choose Properties from the Options menu, and pick Formula from the Type menu. (Adobe: Edit document structure with the Content and Tags panels)

What Taggart does

Taggart applies a heuristic and reports what it finds as a warning, with the line’s position on the page:

Page 1 contains a line dense in mathematical operators but the page has no <Formula> element.

The test is narrow, and it is better to know exactly how narrow. Taggart looks at each line of text between 4 and 200 characters long, on pages that have no Formula element at all. A line is reported when it holds at least two characters from a fixed list, and those make up more than 8% of its non-space characters. The list is ∑ ∏ ∫ √ ± × ÷ ≠ ≤ ≥ ≈ ∞ ∂ ∇ ∈ ∉ ⊂ ⊃ ∪ ∩ and the Greek letters α β γ δ ε θ λ μ π σ τ φ χ ψ ω Γ Δ Θ Λ Ξ Π Σ Φ Ψ Ω. It reports the first such line on a page and moves on.

What that means in practice:

  • It catches text equations tagged as P, and also those tagged as Figure.
  • It does not catch a² + b² = c²: plus, equals and superscript digits are not on the list.
  • A page with one Formula is skipped entirely, so a second expression on the same page tagged as a paragraph is not reported.
  • An expression drawn as a picture has no text for the heuristic to read.

There is no repair. Where an expression begins and ends is the decision the condition is about, so the tagging is yours, in Acrobat or at the source.

How Taggart handles it

The protocol marks this condition as needing human judgement, so no validator can decide it outright. Taggart applies a heuristic to find the likely cases and reports each one as warning for you to settle in the review queue. A heuristic finds what it was built to find and can miss the rest; what this one looks for is described above.

Whatever Taggart changes, the page still looks identical. Fixes edit the PDF object model, not the content streams, and every remediation is re-rendered and compared against the original at structural similarity ≥ 0.999 before you get the file back. A tool that quietly reflows your document is worse than no tool.

Frequently asked questions

Does inline maths count, or only displayed equations?

Both. PDF/UA-1 says all mathematical expressions, so p ≤ 0.05 in the middle of a sentence is a Formula as much as a displayed equation is. Where the line falls is the judgement: "a 2 × 4 grid" in prose is not an expression anybody needs read as one.

My equations are images. Is tagging them as Figure enough?

No. A mathematical expression belongs in a Formula whatever it is drawn with. The Figure type with good alt text gets the words to a listener, but it tells software the expression is a picture, and PDF/UA-1 asks for Formula.

Why does Taggart only warn?

Because deciding that a run of characters is a mathematical expression is the judgement the protocol leaves to a person. Taggart's test is a heuristic over the text on the page; it finds likely cases and misses others, and the section below says which.

What happens after I tag it as Formula?

It needs /Alt, the expression in words — that is condition 17-002, and it is a blocker. Its characters also need to map to Unicode, which is 17-003.

Where this sits in the standards

Standards this condition maps to
StandardReference
Matterhorn Protocol 1.1Checkpoint 17 (Mathematical Expressions), index 17-001
PDF/UA-1 (ISO 14289-1)Clause 7.7-1
WCAG 2.11.1.1 Non-text Content — Level A
EN 301 549 / Section 508 / ADA Title IIAll three point at WCAG 2.1 Level AA for non-web documents, so a Level A or AA criterion here is in scope for each of them.

Other conditions in checkpoint 17

Checkpoint 17 covers mathematical expressions. These are its other failure conditions; the ones without a link are in the rule set but do not have a written page yet.

Check your own file. Taggart’s validator is free and unlimited — 75 of the 87 Matterhorn conditions software can decide, with the other 12 listed for you to confirm; no watermark, no expiry, and no account needed up to 60 pages.

Check a PDF free

Checked with Taggart on a generated file that fails this condition on . The Acrobat steps follow Adobe’s documentation and were not walked through by hand. Last updated .