Do robots follow links on a page which uses rel=”canonical” to point to a different page?

Matt's answer:

MATT CUTTS: Today’s question is a little bit of a confusing one. But I’ll try to tackle it. It’s from Ad in Paris who asks on a website, if a page B implements a rel=”canonical” tag with URL of page A, do robots follow every link in page B? So essentially this person said OK, here’s page B and it’s using rel=”canonical” to say oh I’m actually page A. Do you follow all the outgoing links? Well, whether we follow links is independent of whether these two pages are considered the same page. So if there’s enough page rank on this page, we’ll typically follow the outgoing links and find anything that it links to. But rel=”canonical” is somewhat independent or orthogonal of that. In the sense that we’ll take the URL on page B, and then you say rel=”canonical” is really here, so that might map it to the new location, but then we’re following the outgoing links, and we’ll just follow those wherever they lead to. So I think that they’re pretty much separate. Whether we follow links tends to depend on how much page rank a particular page has. If we just barely got to a page at the very end of our crawl, then we might see the outgoing links, but we might not get a chance to follow them and process them. And at some point you always have to sort of stop the crawl, and start again, and sort of start refreshing your index. So there will always be at least a few pages on the web that we haven’t been able to actually go and fetch those pages. So if you do have relatively high page rank, that’s one of the easy ways where we’ll typically follow all the links to go out from that page.

by Matt Cutts - Google's Head of Search Quality Team

Original video: