Selected answers from the Dumb SEO Questions G+ community.
Scott Clark: So that when the posts are shared, authorship and other graph data is preserved. It makes for well constructed social posts that inherit the meta data of the original.
Neil Cheesman: Okay... but why stating that the `Author` is the Facebook page and not the website - for example? Could that be `diluting` the authority of the website in favour of the facebook page?
Scott Clark: It`s really meant to be individual authors if I understand the Author tagging.
Neil Cheesman: But, in the case of a single `author` website - ie the content is from the same person.
Scott Clark: Author tagging can enhance E-A-T signals for evaluating content, also Expertise, Authority and Trustworthiness.
Neil Cheesman: Yes, I can see that - but why make the Facebook Page as the author - why not retain it as the website?
Neil Cheesman: I guess the Google+ publisher/author tags will soon vanish with G+ demise...
Scott Clark: Not really. That portion of G+ will live on. The knowledge graph`s topic layer and other new parts of Google`s Machine Learning can utilize EAT/Author signals to strengthen connections between author nodes and topic nodes, etc. edit: discovery->topicGoogle has said they don`t rank sites based on author reputation, but elsewhere there is discussion about author trust in the quality rater guidelines. Since their goal is machine learning, I can`t see how author reputation can stay out of it. My 2c.
Scott Clark: I don`t think Google fired Othar Hansen ("Authorship Czar") but I think he`s probably at work somewhere on some version of this tied directly to machine learning. Google keeps talking about identifying authors - whatever they call it - in dribs and drabs and it`s a pretty chunky part of the Quality Rater Guidelines. I think that Entity Ids DO exist for authors just like anything in the graph and so far as their trustworthy (social proof) I think they will play into post rank and future post rank for an author. Gary Illyes posted (on whether authorship is still in use) "we are not using authorship at all anymore…we are smarter than that" I read that as a nod to machine learning. I could be way off.
David Ogletree: Scott Clark I just remember there was a short time it was real important and then they changed how they used it.
Scott Clark: David Ogletree you`re correct. I think they realized that doing authorship "explicitly" was incompatible with just using authors as entities and added both overhead and spam-ability to the mix.