deadheadhtml element without a lang
An <html> without a non-empty lang leaves the page's language unknown, so screen readers guess how to pronounce it; WCAG 3.1.1, Level A.
- Avoid
- Harmful
- Severity
- HarmfulActively breaks something for users.
- Basis
- SpecifiedThe evidence is in a current standard.
- Detectable
- YesMatched exactly. Autofixable when the rule carries a fix.
- Matches
html + logic moduleElement rule, anywhere in the document.- Fix
- None — reports only
- Impacts
- Accessibility and SEO
- Related
head/viewport
<html> with no lang, or with lang="", doesn’t say what language the page is written
in. Browsers and assistive technology still have to render and read it in some language,
so they fall back to a guess, usually the user’s own default, whatever the page actually
contains.
Why avoid
The HTML Standard makes the gap explicit. lang takes “a valid BCP 47 language tag, or
the empty string”, and the empty string means the language is unknown. With no lang
on any ancestor, the language comes from a Content-Language pragma or header if there is
one, and otherwise stays unknown. The Standard’s advice is direct: “authors should specify
the lang attribute on the root html element”.
That makes it a Level A accessibility failure. WCAG 2.2 SC 3.1.1 Language of Page requires that “the default human language of each web page can be programmatically determined”, so that “screen readers can load the correct pronunciation rules”. Without it, a screen reader reads the page with the user’s default voice. A French page read by English speech rules is barely intelligible, and a braille display gets the wrong contractions.
A header doesn’t make up for it. The W3C ACT rule for this criterion accepts only a lang
attribute that is “neither empty nor only ASCII whitespace”. Language set through HTTP
headers or <meta> is “not supported by all assistive technologies”, and xml:lang on its
own fails. The same missing value degrades everything else that keys off language:
browser translation offers, spell-checking in form fields, hyphens: auto, and font
selection for CJK text.
Use instead
<!doctype html>
<html lang="en">Use a BCP 47 tag, like en, en-GB, de-CH or zh-Hant. Mark any passage in a different
language on its own element: <span lang="fr">c'est la vie</span>.
Detectability
Fully detectable. The selector finds <html>, and the logic reports it when lang is
missing, empty or only whitespace. A selector alone can’t express the whitespace case.
Whether a present value is a valid language tag isn’t checked here.
Like head/viewport, nothing is reported unless the document has a <head> with at least
one element in it, so fragments, partials and component templates don’t fire. A document
that only ever loads inside an <iframe> is still reported, because a file on disk can’t
know it will be framed. Mark that case with <!-- deadhead-disable document/html-lang -->.
There is no autofix: the value is the page’s language, which only the author knows.
Resources
- HTML Standard — the lang and xml:lang attributes — the empty string means unknown; the fallback chain; “authors should specify the lang attribute on the root html element”.
- W3C — Understanding WCAG 2.2 SC 3.1.1: Language of Page — the Level A criterion, pronunciation rules for screen readers, and technique H57.
- W3C ACT rule b5c3f8 — HTML page has lang attribute —
langmust be neither empty nor only whitespace; headers,<meta>andxml:langalone don’t satisfy it.
This page is generated from content/rules/document/html-lang.md, the same file the linter is built from. Think the rule is wrong, or that browsers moved on? Say so — that is the most useful issue you can file.