When determining the language of a HTML page, I understand that Google looks at any language declarations that the page owner has set, and then also applies its own language detection algorithms.
But does Google similarly look at language meta data set in PDF files when determining a PDF file's language? (Authors of PDF files can set document-wide properties describing the language (or languages) contained within it.)
Or does Google rely exclusively on language detection algorithms and disregard the language flag set within the PDF file?
Can anyone shed any light?