How does Google determine duplicate content when it's translated into different languages?

Question

I was wondering how duplicate content in another language works.

Google requests that you use canonical links to determine duplicate content or duplicates of your pages to avoid duplicate content in your website. I'm wondering what are the effects of this when using other languages?

For instance, is it considered to be duplicate content if the site is in another language which happens to be an exact duplicate? Does Google translate all text into one base language and then check for duplicate content?

I've read that when publishing translations, you should use hreflang tags to specify the language to help Google to pinpoint the right location and serve the correct content to the right users. However, I've also read you are to use rel="alternate" hreflang tags. Is this to tell Google that it is duplicate content, but in a different language?

Answer 1

...is it considered to be duplicate content if the site is in another language which happens to be an exact duplicate?

Google doesn't consider the same content translated into different languages as duplicate content since the same content in English that's translated into French is different, unlike the same content appearing twice in English, as covered by Matt Cutts here: Does translated content cause a duplicate content issue?

Does Google translate all text into one base language and then check for duplicate content?

Although Google doesn't consider a translation to another language as duplicate content, it does matter how you translate the content:

If you used Google Translate to preform the translation, then the Googlebot would know that since auto-translation embeds various Meta tags, JavaScript, and CSS into the translated page, along with concatenating its translation URLs into all of the links so that the linked pages will also appear translated.

Thus by checking for various translation clues, the Googlebot can tell if Google performed the translation, and to/from what languages. If detected, then it's just a matter of reversing the To and From language pairs to translate it back into the original language in order to make the comparison.

Google prefers human translations instead that are tailored to the locale and audience, as Matt Cutts discusses here (~0:41), and might view lots of auto-translated pages as spammy.

I've also read you are to use rel="alternate" hreflang tags. Is this to tell Google that it is duplicate content, but in a different language?

Exactly. As covered by Google in regards to using hreflang: This markup tells Google's algorithm to consider all of these pages as alternate versions of each other. So you won't need to be concerned about duplicate content issues when that's used.

Answer 2

The purpose of hreflang is to help Google with geotargeting your website, especially if you are having all your language versions in different subfolders on a gTLD domain (like .com, .net, .org etc.). So its main purpose is to serve the correct version of your website to the user, depending on their language and region. It can also help prevent from duplicate content issues if you are targeting different regions that use the same language, such as English speaking regions like UK, US, CA, and so on.

As for the base language of your homepage, you can define this for yourself. Big brands might choose to set English as the default language for their .com-site, but that might as well depend on the origin of the brand (eg a French brand might choose French as their default language). You don't necessarily have to set a default language though.

Answer 3

Regarding the hreflang tag:

By using this tag, you are telling Google that the content on both websites is intended for different audiences and is not duplicate content.

Whether Google will flag duplicate content:

If you just translate the text you may still be penalized. To prevent this you can change the wording, tone, and sentence structure to make it unique.