(1)
A Google Groups thread shares an interesting insight in how Google may interpret a site's language.
A webmaster asked Google why a site with no pages in Spanish is being interpreted as a Spanish site in Google. The site in question is vsanswers.com, and if you look through all the HTML pages, none of them have any Spanish words. But if you look at some of the PDF files on the site, you will discover Spanish in those documents.
Googler, JohnMu, explained:
Looking at some of the pages indexed for your site, it appears that you have some PDFs in multiple languages.I assume that some of the Spanish keywords can be found on pages like those. In general, PDFs like that can be a bit problematic as it's nearly impossible to determine the language they're in, which could result in the document ranking for an interesting mixture of keywords. If that's ok with you, then you can certainly leave it -- otherwise you could use your robots.txt file to prevent crawling of these files, making them drop out of the index over time.
It appears to me that the webmaster has removed the PDFs from the Google index and that Google may soon resolve the site's issue.
I do find it interesting that while the whole site is in English, that a few PDF documents with some Spanish in them would cause such an issue? To be honest, JohnMu's response is a little confusing to me. So maybe I am missing something and maybe people can clarify in the comments. John said, "PDFs like that can be a bit problematic as it's nearly impossible to determine the language they're in," if that is the case, how can Google qualify the site as being Spanish, if they have a hard time determining the language in the PDF? Or maybe John means that Google misunderstood the PDFs as being in Spanish, which caused the issue? But if that is the case, i.e. Google has a hard time determining the language in the PDF, then why would Google use PDFs in the case of determining the language of the site - why not just use the HTML pages?
Forum discussion at Google Groups.
