This is a problem with a lot of paginated sites, such as Tumblr, various forums, and comment pages. Anything that's ordered from newest to oldest won't have a constant correspondence between URL and content.
Google should give the actual article URL a higher score and pagination pages a lower score. So that in their search results I see the content first and the "dupe content" on pagination pages not at all (or way down). (at least for common blog software)
Now imagine that Google links to example.com/?page=2 as it found the search phrase also there (at a given time only Google knows). So when the user clicks on the search result link that leads to example.com/?page=2 NOW should the blog software know what Google or the user wants?
One thing that comes to my mind is to use the referrer and if it's a common search engine parse the s=SEARCHTERM string and use an internal article search to find the best matching article.