Users expect default quotation marks around quotes to conform to the conventions of a particular language. For example, en-US typically uses “ ...”, whereas in Greek text a user may expect to see «...».

In a multilingual document, the quotation marks are part of the surrounding context, and don’t depend on the language of the quoted text (I've omitted the final period of the sentence, since the location of the period is irrelevant to the issue at hand):

But Lucy replied: “Embrassez George de ma part”

If we were to add the subquote, publishers would expect to see that use English quotation marks too, unless we specifically wanted to reproduce the text being quoted exactly as it looked in the original. So we’d normally expect to see:

But Lucy replied: “Embrassez George de ma part et dites-lui, ‘Embrouille’”

instead of French quotation marks.

The i18n WG would like to propose that the version of the quotation marks should be set according to the language indicated on the html tag, and not according to the language of the quoted text.

Currently, there are two sets of tests produced by the i18n WG:

1. https://w3c.github.io/i18n-tests/results/the-q-element.html
2. https://www.w3.org/International/tests/repo/results/the-q-element.en.html

The first one tests the proposed behavior above, and the second one tests the behavior/rendering currently specified in the WHATWG HTML Standard.

It seems that there're no tests about <q> in WPT at the moment:

• https://github.com/w3c/web-platform-tests/tree/master/html/semantics/text-level-semantics/the-q-element
• https://github.com/w3c/web-platform-tests/tree/master/html/rendering/non-replaced-elements/quotes

I'd be happy to port one of the tests above to WPT, preferably the first one.

For more information:

• https://w3c.github.io/i18n-drafts/questions/qa-the-q-element.en
• https://www.w3.org/2018/04/19-i18n-minutes.html#item03