This is not the final website. It is a demo of the real tool working, made for Y Combinator.
AxisRemedy

11-001 Natural language for text in page content cannot be determined.

The natural language of the page text cannot be determined.

Checkpoint
11: Declared Natural Language
ISO 14289-1 (PDF/UA-1) clause
7.2-3
Test type in the protocol
Machine-testable
Applies to
A single object
WCAG 2.x (AxisRemedy's informative mapping)
3.1.1 Language of Page (level A)

How to fix it

The catalog of the file has no usable Lang entry, so a screen reader does not know which language to pronounce the text in. A reader tuned for English will read German or Urdu as if it were English. Setting the language once for the whole document fixes it for every page that has no language of its own.

In Acrobat Pro

  1. Open File, then Properties, then the Advanced tab.
  2. Under Reading Options, choose the main language of the document in the Language list.
  3. Save the file.

In Word, before you export

Select all text, open Review, then Language, then Set Proofing Language, and choose the language of the document. Word writes it into the PDF when you export.

What to enter

Source: W3C, Technique PDF16, Setting the default language using the Lang entry in the document catalog of a PDF document. Menu names differ a little between versions.

How AxisRemedy handles it

Tested by the app. AxisRemedy tests this condition itself. For each document it records passed, failed or not applicable, and for every failure the object and the page.

What the check covers

Checks only the catalog /Lang: present, not blank, shaped like a BCP 47 tag. Page content is not parsed, so a language set on individual elements is neither credited nor judged.

Repairs that target it

Repairs are optional. The few that could change how a page looks, such as embedding a missing font, need your explicit consent.