Does Google Parse The Anchor Text while Indexing

[atakala]

Hey moz fanz,

I'm here to ask a bit technical and open-minding question.
In the Google's paper http://infolab.stanford.edu/~backrub/google.html 
They say they parse the page into hits which is basically word occurences.
But I want to know that they also do the same thing while keeping the anchor text database.
I mean do they parse the anchor text or keep it as it is .
For example, let's say my anchor text is "real car games".
When they indexing my link with anchor text, do they parse my anchor text as hits like
"real" distinct hits
"car" distinct hits
"games" distinct hits.
OR do they just use it as it is. As "real car games"

[SamuelScott]

I would say it depends on whether an entity is detected.

Imagine there is a company named "Real SEO." Google crawls a website that mentions them. Google sees the word "real" and then the word "seo." Normally, Google would see that "real" is an adjective that is modifying the noun "seo." So normally, this would be viewed as two separate, distinct words.

However, in this example, "real seo" is a brand and an "entity." So, even though the two words are first viewed separately, Google has become smart enough to figure out that when those two separate words are found in that order, then they are together referring to a single "thing."

For more on entities in search, I'd read the Moz posts here, here, here, and here.