Pronunciation

What it asks

When a word’s meaning is ambiguous without its pronunciation, the page must provide a way for users to know the intended pronunciation. The classic English examples are heteronyms: “lead” (the metal vs. to guide), “bow” (a knot vs. to bend), “read” (present vs. past tense), “row” (a line vs. an argument).

In Japanese, Chinese, and other languages with extensive homography, the issue is much wider — written kanji or hanzi can map to multiple readings with different meanings.

How to meet it

• Provide a phonetic spelling in parentheses next to the ambiguous word: “He took the lead (leed) role.”
• Use ruby annotations (<ruby> and <rt>) for East Asian languages to mark pronunciation above the base text.
• Link to an audio recording of the word.
• Use IPA (International Phonetic Alphabet) where the audience can be expected to read it.
• For poetry and song lyrics where meter relies on a specific pronunciation, annotate the contested words.

Common failures

• Educational content about poetry, music, or linguistics that hinges on pronunciation but provides no pronunciation cues.
• Bilingual content with kanji that has multiple readings, no furigana, and the wrong reading would change meaning.
• Brand names, place names, or scientific terms with non-obvious pronunciation given without any guide.

Why it matters

This is a niche AAA criterion that matters most in education, linguistics, language-learning sites, and East Asian content. Screen readers can mispronounce ambiguous heteronyms regardless of any markup — the SC asks that human readers also have a way to disambiguate, since the wrong pronunciation can flip the meaning of a sentence. The technique is editorial: when you write “She wound the bandage around the wound,” that needs annotation.