Just to clarify what I mean, I have seen some review articles that have bibliography items like this:
Bloggs, J., "The Paper that has Been Written", The Journal Title, Series Number, page number(s), (year). [DOI] [arXiv link]
where both [DOI] and [arXiv link] are hyperlinks to the paper, the former to its webpage on the journal website (with which it was published), the latter to its arXiv webpage.
Does the author cite twice just in case someone is reading it who might not have access to journals? Or is there another reason?
I agree with Dave Clarke. I think the primary purpose of a bibliography should be to offer the reader a way to read the cited material. A version of the paper on the arXiv achieves this goal better than any other means I know. However, it's also important that the cited material has been verified. By citing the journal version of the paper, the author communicates that it has been.